Compare commits

..

27 Commits

Author SHA1 Message Date
Fred Navruzov
31129b9d8c docs/vmanomaly: fix outdated anchors and param names (#11064)
Updated /docs/anomaly-detection/ section content for stale anchors, metrics renames, and typos in param names

PR https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11064
2026-06-05 21:45:31 +03:00
Max Kotliar
3147f9c24b go.mod: security update golang.org/x/crypto
Fixes:

NAME                 INSTALLED  FIXED IN  TYPE       VULNERABILITY
SEVERITY  EPSS           RISK
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5006
Critical  < 0.1% (21st)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5023
Critical  < 0.1% (16th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5017
Critical  < 0.1% (17th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5020
Critical  < 0.1% (17th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5013   High
< 0.1% (17th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5005
Critical  < 0.1% (13th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5021
Critical  < 0.1% (11th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5019
Critical  < 0.1% (9th)   < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5018   High
< 0.1% (10th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5033
Medium    < 0.1% (16th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5014
Medium    < 0.1% (10th)  < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5015
Medium    < 0.1% (8th)   < 0.1
golang.org/x/crypto  v0.51.0    0.52.0    go-module  GO-2026-5016
Medium    < 0.1% (6th)   < 0.1
2026-06-05 20:48:57 +03:00
Max Kotliar
db8c9badbf docs/changelog: cut release v1.145.0
Signed-off-by: Max Kotliar <mkotlyar@victoriametrics.com>
2026-06-05 15:42:48 +03:00
Max Kotliar
2dc487d125 app/vmselect: run make vmui-update
Signed-off-by: Max Kotliar <mkotlyar@victoriametrics.com>
2026-06-05 15:34:42 +03:00
Nikolay
4cb5fae347 lib/storage: add scheduled rows metrics for calculation background merge completion (#1048)
Previously, only the size and number of partitions were exposed as metrics. It's not enough to estimate the possible completion time.

New metrics `vm_downsampling_partitions_scheduled_rows` and `vm_retention_filters_partitions_scheduled_rows` allow predicting completion time by the following query:

```
sum(vm_downsampling_partitions_scheduled_rows)
/
sum(rate(vm_rows_merged_total{type="storage/big"}))
```
and

```
sum(vm_retention_filters_partitions_scheduled_rows)
/
sum(rate(vm_rows_merged_total{type="storage/big"}))
```

Note, `vm_rows_merged_total` accounts for all possible merges, and it could slightly overestimate completion time.

Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10960
PR https://github.com/VictoriaMetrics/VictoriaMetrics-enterprise/pull/1048
2026-06-05 14:51:46 +03:00
Max Kotliar
3c192f9238 docs/changelog: add link to issue 2026-06-04 17:47:19 +03:00
Max Kotliar
159bc15825 docs/changelog: add link to pr 2026-06-04 17:41:09 +03:00
Max Kotliar
8db58ac410 docs/changelog: add link to mimir page 2026-06-04 17:38:51 +03:00
Max Kotliar
42c1f729db dashboards: show short_version when available, fall back to long version otherwise (#11047)
The `Version` panel shows an empty value when a service is built from a
feature branch. In such cases, `short_version` label is not available,
so `by(job, short_version)` grouping produces no visible value in the
panel.

Previously, to see the actual version, one had to manually edit the
panel query and add the relevant `version` label to the `by()` clause.

The fix tries `short_version` and when empty falls back to a long
`version` label.

This is applied uniformly across all dashboards:
- victoriametrics (single-node)
- victoriametrics-cluster
- vmagent
- vmalert
- vmauth
- operator

PR simplifies debugging with custom or feature-branch builds.
2026-06-04 17:28:25 +03:00
Max Kotliar
6851a75c71 lib/netutil: detect silently dropped idle TCP connections before reuse (#11024)
Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10735
Related to https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10646

Connections silently dropped at the network layer (no RST/FIN) remain
alive from the OS perspective until TCP_USER_TIMEOUT fires. It happens while the connection stays in the pool. When such a stale connection is reused from the pool, the next write fails immediately (<100 microseconds) with a `write: connection timed out` error. Since timeout errors are not retriable, it results in a user-visible query error like:
```
cannot flush requestData to conn: write tcp4 ...: write: connection timed out
```

Before returning a pooled connection to a caller, probe it with a non-blocking read, but only if it has been idle for 5 or more seconds. Fresh connections (idle < 5s) are returned without an extra alive check to avoid unnecessary overhead on the fast path.

The alive probe sets ReadDeadline=now+5ms and reads 1 byte. A deadline-exceeded error means the connection is alive. Any other result (EOF, ETIMEDOUT, unexpected data) means the connection is dead, and it is discarded.

PR https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11024

This is useful for deployments with unstable network infrastructure,
where established connections can be silently broken while sitting in
the pool.

When in doubt if it's your case or not, look for `write: connection
timed out` errors in logs, or monitor the error ratio with:

```
sum(
  increase(vm_request_errors_total{action="search",type="rpcClient",name="vmselect"}[1m])
) by (job)
/
sum(
  increase(vm_requests_total{action="search",type="rpcClient",name="vmselect"}[1m])
) by (job)
* 100
```
2026-06-04 17:13:44 +03:00
JAYICE
1baaaf3b31 app/vmagent: introduce metrics for kafka saturation breakdown
This commit adds new metrics for kafka remotewrite:

* vmagent_remotewrite_kafka_outbuf_latency_seconds
* vmagent_remotewrite_kafka_rtt_seconds

 It helps to detect possible network saturation between vmagent and kafka brokers.

fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10730
2026-06-04 13:28:05 +02:00
Zhu Jiekun
ec88b9cac6 app/vmauth: properly log user information when a missing route error occurs
properly log user information when a missing route error occurs.

examples:
- `user foo missing route for "http://foo:secret@some-host.com/a/b`
- `user unauthorized missing route for "http://some-host.com/abc?de=fg"`

---

Previously, the error log was not good enough to identify the source of
the request:
```
2026-06-03T06:48:09.020Z	warn	VictoriaMetrics/app/vmauth/main.go:427	remoteAddr: "**.**.**.**, X-Forwarded-For: **.**.**.**"; requestURI: /prometheus/api/v1/write; missing route for "/prometheus/api/v1/write" 
```
It takes extra efforts to check it by metrics such as
`vmauth_user_requests_total`.

Related PR https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11052
2026-06-04 13:14:55 +02:00
Max Kotliar
661fbf947c docs/changelog: add empty line
follow-up on prev commit
5c176838d1
2026-06-03 13:26:14 +03:00
JAYICE
5c176838d1 deployment: upgrade golang builder from 1.26.3 to 1.26.4 (#11053)
upgrade golang builder to fix failed PR pipeline
https://github.com/VictoriaMetrics/VictoriaMetrics/actions/runs/26870814471/job/79245400038?pr=11052

PR https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11053
2026-06-03 13:21:38 +03:00
Zakhar Bessarab
9d4c06210c app/vmauth: take JWT users into a total count of users for startup message
Previously, config like this:
```
users:
  - jwt:
      skip_verify: true
    url_prefix:
      - "http://backend"
```
Would print a message saying that 0 users were loaded which leads to
confusion.

Related PR https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11050/
2026-06-02 17:51:11 +02:00
Dmytro Kozlov
347d2e0fef app/vmctl: implement migration from mimir object storage
This commit adds the ability to read blocks from Mimir object storage, process
them, and store them in VictoriaMetrics.
This new version of the migration can read Mimir object storage from the
file system and S3, GCP and Azure.

Under the hood, the vmctl tries to read the bucket-index.json file and
collect blocks via the defined filters. Depending on the used path vmctl
decides how to retrieve data. If it is not the file system, it will
download each file from the block and store it in the `tmp` folder.
Process this block folder and remove it from the file system.

Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/7717
2026-06-02 16:51:54 +02:00
Alexander Frolov
f3de1f4ac7 vmselect: allow partial responses for storage node groups
Previously vmselect returned http.StatusServiceUnavailable instead of a partial response when a single storage group is unavailable, even when partial responses are allowed.

According to the cluster documentation, vmselect should continue serving queries from available vmstorage nodes and mark incomplete responses as partial.

This commit fixes behavior and allows partial respones for grouped storage nodes.

Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11009
2026-06-02 16:17:02 +02:00
f41gh7
a9032ecd1d docs/changelog: sort changelog entries 2026-06-02 15:44:08 +02:00
Nikolay
ce712f0bc9 docs/vmauth: add security recommendations for vmselect multitenant reads
Currently, vmselect uses `OR` logic to filter out tenants for
 `multitenant` requests. And if vmauth is used to enforce tenant
 requests with `extra_label` or `extra_filters`. It's required to set
 `extra_label`, `extra_filters` and `extra_filters[]` explicitly at
 configuration file to prevent possible enforcement bypass.
2026-06-02 15:34:17 +02:00
Nikolay
23a3ff4174 app/vmselect: fixes utf-8 escaping at federate API
Prometheus compatible scrapers failed to parse `/federate` API responses
from VictoriaMetrics, if it contains any utf-8 characters.

This commit adds best effort negotiation check for /federate API
requests. By default, VictoriaMetrics emmits output in the same format.
But in case of Accept: allow-utf-8 header, VictoriaMetrics switches into
quoted output for time series name and label names according to
Prometheus v3.0 text exposition format.
https://prometheus.io/docs/guides/utf8/#querying

Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10968
2026-06-02 14:41:07 +02:00
Hui Wang
b638f95aba app/vmalert: reset group evaluation timestamp when system clock is modified backwards
The issue could happen subtly and be hard to debug. It might also
generate confusing results. For example, if the system clock for vmalert
is moved 30m backward, vmalert would reset the evaluation timestamps to
30m ago, causing duplicate evaluations for timestamps within that 30m
window. Exposing a new group metric `vmalert_iteration_reset_total` to
help debuging this issue.

fixes:
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10985
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10423.
2026-06-02 14:40:15 +02:00
Alexander Frolov
2345b2b4ed vmui: clarify partial results warning text
Update the VMUI partial results warning to say `storage nodes` instead
of `vmstorage nodes`.

People operating VictoriaMetrics and people primarily using the UI are
often different audiences. UI users may not know what `vmstorage` means,
but `storage nodes` is easier to understand while preserving the meaning
of the warning.
2026-06-02 14:39:13 +02:00
Hui Wang
b6edf40198 app/vmalert: support match[] parameter in /api/v1/rules and /api/v1/alerts APIs
If multiple match[] parameters are passed, rules that match any of the
provided label selector sets are returned.
For `/api/v1/rules`, the matching is performed against the labels
defined in each rule configuration;
For `/api/v1/alerts`, the matching is performed against the labels
attached in each alert.

The current `search` in [VMUI
alerting](https://play.victoriametrics.com/select/0/prometheus/graph/#/rules)
can only filter by group and rule name. The label filtering in the
vmalert UI is implemented on the frontend and matches exact label
content rather than label selectors. Both UI could be improved after
support this parameter.

fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11020
2026-06-02 14:38:36 +02:00
Nikolay
b220066049 lib/prompb: properly reuse memory for native histogram parsing
Previously, bucket spans were incorrectly re-used due to missing `clear`
call.
It could produce unexpected results with span bucket value from
previously parsed histogram.
 In addition, if timeseries had a multiple data values for the same
 histogram, metric name could add unexpected `bucket` prefix multiple
 times, producing `_bucket_bucket` metric names.

  This commit properly clears reused memory.

 Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11041
2026-06-02 14:36:58 +02:00
Aliaksandr Valialkin
9ea1770ba4 app/vmctl: mention --vm-significant-figures option in the description for the --vm-round-digits option
These options are related, so it is better from the discoveribility PoV to cross-mention them.
2026-06-01 22:18:11 +02:00
Aliaksandr Valialkin
24d176fd2c app/vmagent: mention the -remoteWrite.significantFigures command-line flag in the description for the -remoteWrite.roundDigits flag
These flags are related, so it is better from the discoverity PoV to cross-mention them.
2026-06-01 22:17:16 +02:00
Hui Wang
e5c277237e app/vmalert: fix notifiers page in web UI appearing blank (#11036)
Fixes https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11035.

There is no search box on the vmalert UI Notifiers page, but the notifier group containers have the `vm-group` CSS class applied. This class is hidden by default and only becomes visible after `filterRules()` adds the `vm-found` class. Since `filterRules()` is only called when a search box exists, it is never invoked on the Notifiers
page, leaving all content invisible.

The bug was introduced in v1.129.0 in 32ed45b672 (diff-e349265135dddcf960e58d2ada6be0fc18b76603f74c05107cbd1f348eb4d62b)

PR: https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11036

Co-authored-by: Max Kotliar <mkotlyar@victoriametrics.com>
2026-06-01 18:25:52 +03:00
74 changed files with 2489 additions and 5416 deletions

View File

@@ -79,7 +79,8 @@ var (
"writing them to remote storage. "+
"Examples: -remoteWrite.roundDigits=2 would round 1.236 to 1.24, while -remoteWrite.roundDigits=-1 would round 126.78 to 130. "+
"By default, digits rounding is disabled. Set it to 100 for disabling it for a particular remote storage. "+
"This option may be used for improving data compression for the stored metrics")
"This option may be used for improving data compression for the stored metrics. "+
"See also -remoteWrite.significantFigures")
sortLabels = flag.Bool("sortLabels", false, `Whether to sort labels for incoming samples before writing them to all the configured remote storage systems. `+
`This may be needed for reducing memory usage at remote storage when the order of labels in incoming samples is random. `+
`For example, if m{k1="v1",k2="v2"} may be sent as m{k2="v2",k1="v1"}`+

View File

@@ -95,6 +95,7 @@ type groupMetrics struct {
iterationTotal *metrics.Counter
iterationDuration *metrics.Summary
iterationMissed *metrics.Counter
iterationReset *metrics.Counter
iterationInterval *metrics.Gauge
}
@@ -330,6 +331,7 @@ func (g *Group) Init() {
g.metrics.iterationTotal = g.metrics.set.NewCounter(fmt.Sprintf(`vmalert_iteration_total{%s}`, labels))
g.metrics.iterationDuration = g.metrics.set.NewSummary(fmt.Sprintf(`vmalert_iteration_duration_seconds{%s}`, labels))
g.metrics.iterationMissed = g.metrics.set.NewCounter(fmt.Sprintf(`vmalert_iteration_missed_total{%s}`, labels))
g.metrics.iterationReset = g.metrics.set.NewCounter(fmt.Sprintf(`vmalert_iteration_reset_total{%s}`, labels))
g.metrics.iterationInterval = g.metrics.set.NewGauge(fmt.Sprintf(`vmalert_iteration_interval_seconds{%s}`, labels), func() float64 {
i := g.Interval.Seconds()
return i
@@ -474,14 +476,16 @@ func (g *Group) Start(ctx context.Context, rw remotewrite.RWClient, rr datasourc
if missed < 0 {
// missed can become < 0 due to irregular delays during evaluation
// which can result in time.Since(evalTS) < g.Interval;
// or the system wall clock was changed backward
missed = 0
// or the system wall clock was changed backward,
// Reset the evalTS to the current time.
evalTS = time.Now()
g.metrics.iterationReset.Inc()
} else {
evalTS = evalTS.Add((missed + 1) * g.Interval)
}
if missed > 0 {
g.metrics.iterationMissed.Inc()
}
evalTS = evalTS.Add((missed + 1) * g.Interval)
eval(evalCtx, evalTS)
}

View File

@@ -11,6 +11,8 @@ import (
"strconv"
"strings"
"github.com/VictoriaMetrics/metricsql"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/config"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/notifier"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/rule"
@@ -160,12 +162,12 @@ func (rh *requestHandler) handler(w http.ResponseWriter, r *http.Request) bool {
case "/vmalert/api/v1/alerts", "/api/v1/alerts":
// path used by Grafana for ng alerting
gf, err := newGroupsFilter(r)
af, err := newAlertsFilter(r)
if err != nil {
errJson(w, r, err)
return true
}
data, err := rh.listAlerts(gf)
data, err := rh.listAlerts(af)
if err != nil {
errJson(w, r, err)
return true
@@ -325,6 +327,48 @@ func (gf *groupsFilter) matches(group *rule.Group) bool {
return true
}
type alertsFilter struct {
gf *groupsFilter
match [][]metricsql.LabelFilter
}
func getMatchFilters(matches []string) ([][]metricsql.LabelFilter, *httpserver.ErrorWithStatusCode) {
if len(matches) == 0 {
return nil, nil
}
tfss := make([][]metricsql.LabelFilter, 0, len(matches))
for _, s := range matches {
expr, err := metricsql.Parse(s)
if err != nil {
return nil, errResponse(fmt.Errorf(`invalid parameter "match[]": failed to parse %q: %w`, s, err), http.StatusBadRequest)
}
me, ok := expr.(*metricsql.MetricExpr)
if !ok {
return nil, errResponse(fmt.Errorf(`invalid parameter "match[]": expecting metricSelector; got %q`, expr.AppendString(nil)), http.StatusBadRequest)
}
if len(me.LabelFilterss) == 0 {
return nil, errResponse(fmt.Errorf(`invalid parameter "match[]": labelFilterss cannot be empty`), http.StatusBadRequest)
}
tfss = append(tfss, me.LabelFilterss...)
}
return tfss, nil
}
func newAlertsFilter(r *http.Request) (*alertsFilter, *httpserver.ErrorWithStatusCode) {
gf, err := newGroupsFilter(r)
if err != nil {
return nil, err
}
var af alertsFilter
af.gf = gf
af.match, err = getMatchFilters(r.Form["match[]"])
if err != nil {
return nil, err
}
return &af, nil
}
// see https://prometheus.io/docs/prometheus/latest/querying/api/#rules
type rulesFilter struct {
gf *groupsFilter
@@ -335,6 +379,7 @@ type rulesFilter struct {
maxGroups int
pageNum int
search string
match [][]metricsql.LabelFilter
extendedStates bool
}
@@ -355,7 +400,10 @@ func newRulesFilter(r *http.Request) (*rulesFilter, *httpserver.ErrorWithStatusC
return nil, errResponse(fmt.Errorf(`invalid parameter "type": not supported value %q`, ruleTypeParam), http.StatusBadRequest)
}
}
rf.match, err = getMatchFilters(r.Form["match[]"])
if err != nil {
return nil, err
}
states := vs["state"]
if len(states) == 0 {
states = vs["filter"]
@@ -416,12 +464,47 @@ func (rf *rulesFilter) matchesRule(r *rule.ApiRule) bool {
if len(rf.ruleNames) > 0 && !slices.Contains(rf.ruleNames, r.Name) {
return false
}
if !areLabelsMatch(r.Labels, rf.match) {
return false
}
if len(rf.states) == 0 {
return true
}
return slices.Contains(rf.states, r.State)
}
func areLabelsMatch(labels map[string]string, matches [][]metricsql.LabelFilter) bool {
if len(matches) == 0 {
return true
}
// labels need to match at least one of the provided match[] arg
return slices.ContainsFunc(matches, func(filters []metricsql.LabelFilter) bool {
for _, mf := range filters {
if !isLabelFilterMatch(labels[mf.Label], mf) {
return false
}
}
return true
})
}
func isLabelFilterMatch(s string, match metricsql.LabelFilter) bool {
if !match.IsRegexp {
if match.IsNegative {
return s != match.Value
}
return s == match.Value
}
re, err := metricsql.CompileRegexpAnchored(match.Value)
if err != nil {
return false
}
if match.IsNegative {
return !re.MatchString(s)
}
return re.MatchString(s)
}
func (rh *requestHandler) groups(rf *rulesFilter) *listGroupsResponse {
rh.m.groupsMu.RLock()
defer rh.m.groupsMu.RUnlock()
@@ -543,14 +626,14 @@ func (rh *requestHandler) groupAlerts() []rule.GroupAlerts {
return gAlerts
}
func (rh *requestHandler) listAlerts(gf *groupsFilter) ([]byte, *httpserver.ErrorWithStatusCode) {
func (rh *requestHandler) listAlerts(af *alertsFilter) ([]byte, *httpserver.ErrorWithStatusCode) {
rh.m.groupsMu.RLock()
defer rh.m.groupsMu.RUnlock()
lr := listAlertsResponse{Status: "success"}
lr.Data.Alerts = make([]*rule.ApiAlert, 0)
for _, group := range rh.m.groups {
if !gf.matches(group) {
if !af.gf.matches(group) {
continue
}
g := group.ToAPI()
@@ -558,7 +641,11 @@ func (rh *requestHandler) listAlerts(gf *groupsFilter) ([]byte, *httpserver.Erro
if r.Type != rule.TypeAlerting {
continue
}
lr.Data.Alerts = append(lr.Data.Alerts, r.Alerts...)
for _, alert := range r.Alerts {
if areLabelsMatch(alert.Labels, af.match) {
lr.Data.Alerts = append(lr.Data.Alerts, alert)
}
}
}
}

View File

@@ -348,7 +348,7 @@
typeK, ns := keys[i], targets[notifier.TargetType(keys[i])]
count := len(ns)
%}
<div class="w-100 flex-column vm-group">
<div class="w-100 flex-column">
<span class="d-flex justify-content-between" id="group-{%s typeK %}">
<a href="#group-{%s typeK %}">{%s typeK %} ({%d count %})</a>
<span
@@ -361,7 +361,7 @@
<div id="item-{%s typeK %}" class="collapse show">
<table class="table table-striped table-hover table-sm">
<thead>
<tr class="vm-item">
<tr>
<th scope="col">Labels</th>
<th scope="col">Address</th>
</tr>

View File

@@ -1115,7 +1115,7 @@ func StreamListTargets(qw422016 *qt422016.Writer, r *http.Request, targets map[n
//line app/vmalert/web.qtpl:350
qw422016.N().S(`
<div class="w-100 flex-column vm-group">
<div class="w-100 flex-column">
<span class="d-flex justify-content-between" id="group-`)
//line app/vmalert/web.qtpl:352
qw422016.E().S(typeK)
@@ -1152,7 +1152,7 @@ func StreamListTargets(qw422016 *qt422016.Writer, r *http.Request, targets map[n
qw422016.N().S(`" class="collapse show">
<table class="table table-striped table-hover table-sm">
<thead>
<tr class="vm-item">
<tr>
<th scope="col">Labels</th>
<th scope="col">Address</th>
</tr>

View File

@@ -10,6 +10,8 @@ import (
"testing"
"time"
"github.com/VictoriaMetrics/metricsql"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/config"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/datasource"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmalert/notifier"
@@ -37,12 +39,14 @@ func TestHandler(t *testing.T) {
Concurrency: 1,
Rules: []config.Rule{
{
ID: 0,
Alert: "alert",
ID: 0,
Alert: "alert",
Labels: map[string]string{"job": "foo"},
},
{
ID: 1,
Record: "record",
Labels: map[string]string{"job": "bar"},
},
},
}, fq, 1*time.Minute, nil)
@@ -128,6 +132,18 @@ func TestHandler(t *testing.T) {
if length := len(lr.Data.Alerts); length != 2 {
t.Fatalf("expected 2 alert got %d", length)
}
lr = listAlertsResponse{}
getResp(t, ts.URL+`/api/v1/alerts?match[]={job="foo"}`, &lr, 200)
if length := len(lr.Data.Alerts); length != 3 {
t.Fatalf("expected 3 alerts got %d", length)
}
lr = listAlertsResponse{}
getResp(t, ts.URL+`/api/v1/alerts?match[]={job="bar"}`, &lr, 200)
if length := len(lr.Data.Alerts); length != 0 {
t.Fatalf("expected 0 alerts got %d", length)
}
})
t.Run("/api/v1/alert?alertID&groupID", func(t *testing.T) {
expAlert := rule.NewAlertAPI(ar, ar.GetAlerts()[0])
@@ -242,6 +258,13 @@ func TestHandler(t *testing.T) {
check("/vmalert/api/v1/rules?datasource_type=graphite", 200, 1, 2)
check("/vmalert/api/v1/rules?datasource_type=graphiti", 400, 0, 0)
// invalid match[] params
check(`/vmalert/api/v1/rules?match[]={job=!"foo"}`, 400, 0, 0)
check(`/vmalert/api/v1/rules?match[]={job="foo"}`, 200, 3, 3)
check(`/vmalert/api/v1/rules?match[]={job="bar"}`, 200, 3, 3)
check(`/vmalert/api/v1/rules?match[]={job="bar"}&match[]={job="foo"}`, 200, 3, 6)
check(`/vmalert/api/v1/rules?match[]={job="barzz"}`, 200, 0, 0)
// no filtering expected due to bad params
check("/api/v1/rules?type=badParam", 400, 0, 0)
check("/api/v1/rules?foo=bar", 200, 3, 6)
@@ -367,3 +390,116 @@ func TestEmptyResponse(t *testing.T) {
}
})
}
func TestMatchesRule(t *testing.T) {
parseMatch := func(t *testing.T, selectors []string) [][]metricsql.LabelFilter {
t.Helper()
var match [][]metricsql.LabelFilter
for _, s := range selectors {
expr, err := metricsql.Parse(s)
if err != nil {
t.Fatalf("failed to parse selector %q: %v", s, err)
}
me, ok := expr.(*metricsql.MetricExpr)
if !ok {
t.Fatalf("expected MetricExpr for %q, got %T", s, expr)
}
match = append(match, me.LabelFilterss...)
}
return match
}
f := func(t *testing.T, selectors []string, labels map[string]string, wantMatch bool) {
t.Helper()
rf := &rulesFilter{
gf: &groupsFilter{},
match: parseMatch(t, selectors),
}
r := &rule.ApiRule{Labels: labels}
got := rf.matchesRule(r)
if got != wantMatch {
t.Fatalf("matchesRule(%v) with selectors %v: got %v, want %v",
labels, selectors, got, wantMatch)
}
}
f(t, nil, map[string]string{"foo": "bar"}, true)
f(t, []string{`{foo="bar"}`}, map[string]string{"foo": "bar"}, true)
f(t, []string{`{foo="bar"}`}, map[string]string{"foo": "baz"}, false)
f(t, []string{`{foo="bar"}`}, map[string]string{"bar": "baz"}, false)
f(t, []string{`{foo=""}`}, map[string]string{"bar": "baz"}, true)
f(t, []string{`{foo!="bar"}`}, map[string]string{"foo": "baz"}, true)
f(t, []string{`{foo!="bar"}`}, map[string]string{"foo": "bar"}, false)
f(t, []string{`{foo=~"bar.*"}`}, map[string]string{"foo": "bar"}, true)
f(t, []string{`{foo=~"bar.*"}`}, map[string]string{"foo": "baz"}, false)
f(t, []string{`{bar=~"baz|bar"}`}, map[string]string{"bar": "baz"}, true)
f(t, []string{`{bar=~"baz|bar"}`}, map[string]string{"bar": "bar"}, true)
f(t, []string{`{bar=~"baz|bar"}`}, map[string]string{"bar": "foo"}, false)
f(t, []string{`{foo!~"bar.*"}`}, map[string]string{"foo": "baz"}, true)
f(t, []string{`{foo!~"bar.*"}`}, map[string]string{"foo": "bar"}, false)
// single match[] with multiple filters
f(t,
[]string{`{job="foo",instance="bar"}`},
map[string]string{"job": "foo", "instance": "bar"},
true,
)
f(t,
[]string{`{job="foo",instance="bar"}`},
map[string]string{"job": "other", "instance": "bar"},
false,
)
f(t,
[]string{`{foo="bar",baz=~"b.*"}`},
map[string]string{"foo": "bar", "baz": "bazinga"},
true,
)
f(t,
[]string{`{foo="bar",baz=~"b.*"}`},
map[string]string{"foo": "other", "baz": "bazinga"},
false,
)
// multiple matches[]
f(t,
[]string{`{foo="bar"}`, `{foo="baz"}`},
map[string]string{"foo": "baz"},
true,
)
f(t,
[]string{`{foo="bar"}`, `{foo="baz"}`},
map[string]string{"foo": "unknown"},
false,
)
f(t,
[]string{`{foo=~"bar.*"}`, `{bar=~"baz.*"}`},
map[string]string{"bar": "bazinga"},
true,
)
f(t,
[]string{`{foo=~"bar.*"}`, `{bar=~"baz.*"}`},
map[string]string{"foo": "bartender"},
true,
)
f(t,
[]string{`{foo=~"bar.*"}`, `{bar=~"baz.*"}`},
map[string]string{"foo": "other", "bar": "other"},
false,
)
f(t,
[]string{`{job="foo",instance="bar"}`, `{foo="bar"}`},
map[string]string{"foo": "bar"},
true,
)
f(t,
[]string{`{job="foo", instance="bar"}`, `{foo="bar"}`},
map[string]string{"instance": "barr", "job": "foo"},
false,
)
}

View File

@@ -889,7 +889,8 @@ func reloadAuthConfig() (bool, error) {
}
mp := authUsers.Load()
logger.Infof("loaded information about %d users from -auth.config=%q", len(*mp), *authConfigPath)
jwtc := jwtAuthCache.Load()
logger.Infof("loaded information about %d users from -auth.config=%q", len(*mp)+len(jwtc.users), *authConfigPath)
return true, nil
}

View File

@@ -317,7 +317,7 @@ func processUserRequest(w http.ResponseWriter, r *http.Request, ui *UserInfo, tk
defer ui.endConcurrencyLimit()
// Process the request.
processRequest(w, r, ui, tkn)
processRequest(w, r, ui, tkn, userName)
}
func beginConcurrencyLimit(ctx context.Context) error {
@@ -391,7 +391,7 @@ func bufferRequestBody(ctx context.Context, r io.ReadCloser, userName string) (i
return bb, nil
}
func processRequest(w http.ResponseWriter, r *http.Request, ui *UserInfo, tkn *jwt.Token) {
func processRequest(w http.ResponseWriter, r *http.Request, ui *UserInfo, tkn *jwt.Token, userName string) {
u := normalizeURL(r.URL)
up, hc := ui.getURLPrefixAndHeaders(u, r.Host, r.Header)
isDefault := false
@@ -409,7 +409,7 @@ func processRequest(w http.ResponseWriter, r *http.Request, ui *UserInfo, tkn *j
if ui.DumpRequestOnErrors {
di = debugInfo(u, r)
}
httpserver.Errorf(w, r, "missing route for %q%s", u.String(), di)
httpserver.Errorf(w, r, "user %s missing route for %q%s", userName, u.String(), di)
return
}
up, hc = ui.DefaultURL, ui.HeadersConf
@@ -455,7 +455,7 @@ func processRequest(w http.ResponseWriter, r *http.Request, ui *UserInfo, tkn *j
ui.backendErrors.Inc()
}
err := &httpserver.ErrorWithStatusCode{
Err: fmt.Errorf("all the %d backends for the user %q are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend", up.getBackendsCount(), ui.name()),
Err: fmt.Errorf("all the %d backends for the user %q are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend", up.getBackendsCount(), userName),
StatusCode: http.StatusBadGateway,
}
httpserver.Errorf(w, r, "%s", err)

View File

@@ -307,6 +307,24 @@ statusCode=200
requested_url={BACKEND}/bar/a/b`
f(cfgStr, requestURL, backendHandler, responseExpected)
// correct authorization but unexisted path, hence missing route error.
cfgStr = `
users:
- username: foo
password: secret
url_map:
- src_paths:
- "/api/v1/write"
url_prefix: "{BACKEND}/bar"`
requestURL = "http://foo:secret@some-host.com/a/b"
backendHandler = func(w http.ResponseWriter, r *http.Request) {
fmt.Fprintf(w, "requested_url=http://%s%s", r.Host, r.URL)
}
responseExpected = `
statusCode=400
user foo missing route for "http://foo:secret@some-host.com/a/b"`
f(cfgStr, requestURL, backendHandler, responseExpected)
// verify how path cleanup works
cfgStr = `
unauthorized_user:
@@ -403,7 +421,7 @@ unauthorized_user:
}
responseExpected = `
statusCode=400
missing route for "http://some-host.com/abc?de=fg"`
user unauthorized missing route for "http://some-host.com/abc?de=fg"`
f(cfgStr, requestURL, backendHandler, responseExpected)
// missing default_url and default url_prefix for unauthorized user with dump_request_on_errors enabled
@@ -419,7 +437,7 @@ unauthorized_user:
}
responseExpected = `
statusCode=400
missing route for "http://some-host.com/abc?de=fg" (host: "some-host.com"; path: "/abc"; args: "de=fg"; headers:Connection: Some-Header,Other-Header
user unauthorized missing route for "http://some-host.com/abc?de=fg" (host: "some-host.com"; path: "/abc"; args: "de=fg"; headers:Connection: Some-Header,Other-Header
Pass-Header: abc
Some-Header: foobar
X-Forwarded-For: 12.34.56.78
@@ -461,7 +479,7 @@ unauthorized_user:
}
responseExpected = `
statusCode=502
all the 2 backends for the user "" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
all the 2 backends for the user "unauthorized" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
f(cfgStr, requestURL, backendHandler, responseExpected)
// all the backend_urls are unavailable for authorized user
@@ -501,7 +519,7 @@ unauthorized_user:
}
responseExpected = `
statusCode=502
all the 0 backends for the user "" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
all the 0 backends for the user "unauthorized" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
f(cfgStr, requestURL, backendHandler, responseExpected)
netutil.Resolver = origResolver
@@ -518,7 +536,7 @@ unauthorized_user:
}
responseExpected = `
statusCode=502
all the 2 backends for the user "" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
all the 2 backends for the user "unauthorized" are unavailable for proxying the request - check previous WARN logs to see the exact error for each failed backend`
f(cfgStr, requestURL, backendHandler, responseExpected)
if n := retries.Load(); n != 2 {
t.Fatalf("unexpected number of retries; got %d; want 2", n)
@@ -545,6 +563,31 @@ requested_url={BACKEND}/path2/foo/?de=fg`
if n := retries.Load(); n != 2 {
t.Fatalf("unexpected number of retries; got %d; want 2", n)
}
// make sure that empty config value erases client extra filters and extra labels
cfgStr = `
unauthorized_user:
url_prefix: {BACKEND}/foo?bar=baz&extra_filters[]=&extra_label=&extra_filters=`
requestURL = "http://some-host.com/abc/def?some_arg=some_value&extra_filters[]=baz&extra_label=tenant=admin&extra_filters=bar"
backendHandler = func(w http.ResponseWriter, r *http.Request) {
h := w.Header()
h.Set("Connection", "close")
h.Set("Foo", "bar")
var bb bytes.Buffer
if err := r.Header.Write(&bb); err != nil {
panic(fmt.Errorf("unexpected error when marshaling headers: %w", err))
}
fmt.Fprintf(w, "requested_url=http://%s%s\n%s", r.Host, r.URL, bb.String())
}
responseExpected = `
statusCode=200
Foo: bar
requested_url={BACKEND}/foo/abc/def?bar=baz&extra_filters=&extra_filters%5B%5D=&extra_label=&some_arg=some_value
Pass-Header: abc
User-Agent: vmauth
X-Forwarded-For: 12.34.56.78, 42.2.3.84`
f(cfgStr, requestURL, backendHandler, responseExpected)
}
func TestJWTRequestHandler(t *testing.T) {

View File

@@ -146,7 +146,8 @@ var (
Name: vmRoundDigits,
Value: 100,
Usage: "Round metric values to the given number of decimal digits after the point. " +
"This option may be used for increasing on-disk compression level for the stored metrics",
"This option may be used for increasing on-disk compression level for the stored metrics. " +
"See also --vm-significant-figures option",
},
&cli.StringSliceFlag{
Name: vmExtraLabel,
@@ -500,6 +501,96 @@ var (
}
)
const (
mimirPath = "mimir-path"
mimirTenantID = "mimir-tenant-id"
mimirConcurrency = "mimir-concurrency"
mimirFilterTimeStart = "mimir-filter-time-start"
mimirFilterTimeEnd = "mimir-filter-time-end"
mimirFilterLabel = "mimir-filter-label"
mimirFilterLabelValue = "mimir-filter-label-value"
mimirCredsFilePath = "mimir-creds-file-path"
mimirConfigFilePath = "mimir-config-file-path"
mimirConfigProfile = "mimir-config-profile"
mimirCustomS3Endpoint = "mimir-custom-s3-endpoint"
mimirS3ForcePathStyle = "mimir-s3-force-path-style"
mimirS3TLSInsecureSkipVerify = "mimir-s3-tls-insecure-skip-verify"
mimirSSEKMSKeyID = "mimir-s3-sse-kms-key-id"
mimirSSEAlgorithm = "mimir-s3-sse-algorithm"
)
var (
mimirFlags = []cli.Flag{
&cli.StringFlag{
Name: mimirPath,
Usage: "Path to Mimir storage bucket or local folder.",
Required: true,
},
&cli.StringFlag{
Name: mimirTenantID,
Usage: "Tenant ID for Mimir storage",
},
&cli.IntFlag{
Name: mimirConcurrency,
Usage: "Number of concurrently running block readers",
Value: 1,
},
&cli.StringFlag{
Name: mimirFilterTimeStart,
Usage: "The time filter in RFC3339 format to select timeseries with timestamp equal or higher than provided value. E.g. '2020-01-01T20:07:00Z'",
Required: true,
},
&cli.StringFlag{
Name: mimirFilterTimeEnd,
Usage: "The time filter in RFC3339 format to select timeseries with timestamp equal or lower than provided value. E.g. '2020-01-01T20:07:00Z'",
Required: true,
},
&cli.StringFlag{
Name: mimirFilterLabel,
Usage: "Mimir label name to filter timeseries by. E.g. '__name__' will filter timeseries by name.",
},
&cli.StringFlag{
Name: mimirFilterLabelValue,
Usage: fmt.Sprintf("Regular expression to filter label from %q flag.", mimirFilterLabel),
Value: ".*",
},
&cli.StringFlag{
Name: mimirCredsFilePath,
Usage: "Path to file with GCS or S3 credentials. Credentials are loaded from default locations if not set. See https://cloud.google.com/iam/docs/creating-managing-service-account-keys and https://docs.aws.amazon.com/general/latest/gr/aws-security-credentials.html",
},
&cli.StringFlag{
Name: mimirConfigFilePath,
Usage: "Path to file with S3 configs. Configs are loaded from default location if not set. See https://docs.aws.amazon.com/general/latest/gr/aws-security-credentials.html",
},
&cli.StringFlag{
Name: mimirConfigProfile,
Usage: "Profile name for S3 configs. If no set, the value of the environment variable will be loaded (AWS_PROFILE or AWS_DEFAULT_PROFILE), or if both not set, DefaultSharedConfigProfile is used",
},
&cli.StringFlag{
Name: mimirCustomS3Endpoint,
Usage: "Custom S3 endpoint for use with S3-compatible storages (e.g. MinIO). S3 is used if not set",
},
&cli.BoolFlag{
Name: mimirS3ForcePathStyle,
Usage: "Prefixing endpoint with bucket name when set false, true by default.",
Value: true,
},
&cli.BoolFlag{
Name: mimirS3TLSInsecureSkipVerify,
Usage: "Whether to skip TLS verification when connecting to the S3 endpoint.",
},
&cli.StringFlag{
Name: mimirSSEKMSKeyID,
Usage: "SSE KMS Key ID for use with S3-compatible storages.",
},
&cli.StringFlag{
Name: mimirSSEAlgorithm,
Usage: "SSE algorithm for use with S3-compatible storages.",
},
}
)
const (
vmNativeFilterMatch = "vm-native-filter-match"
vmNativeFilterTimeStart = "vm-native-filter-time-start"

View File

@@ -18,6 +18,7 @@ import (
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/auth"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/backoff"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/barpool"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/mimir"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/native"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/remoteread"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/logger"
@@ -297,12 +298,12 @@ func main() {
},
},
{
Name: "thanos",
Usage: "Migrate time series from Thanos blocks (supports raw and downsampled data)",
Flags: mergeFlags(globalFlags, thanosFlags, vmFlags),
Name: "mimir",
Usage: "Migrate time series from Mimir object storage or local filesystem",
Flags: mergeFlags(globalFlags, mimirFlags, vmFlags),
Before: beforeFn,
Action: func(c *cli.Context) error {
fmt.Println("Thanos import mode")
fmt.Println("Mimir import mode")
vmCfg, err := initConfigVM(c)
if err != nil {
@@ -314,6 +315,54 @@ func main() {
return fmt.Errorf("failed to create VM importer: %s", err)
}
mCfg := mimir.Config{
Filter: mimir.Filter{
TimeMin: c.String(mimirFilterTimeStart),
TimeMax: c.String(mimirFilterTimeEnd),
Label: c.String(mimirFilterLabel),
LabelValue: c.String(mimirFilterLabelValue),
},
Path: c.String(mimirPath),
TenantID: c.String(mimirTenantID),
CredsFilePath: c.String(mimirCredsFilePath),
ConfigFilePath: c.String(mimirConfigFilePath),
ConfigProfile: c.String(mimirConfigProfile),
CustomS3Endpoint: c.String(mimirCustomS3Endpoint),
S3ForcePathStyle: c.Bool(mimirS3ForcePathStyle),
S3TLSInsecureSkipVerify: c.Bool(mimirS3TLSInsecureSkipVerify),
SSEKMSKeyID: c.String(mimirSSEKMSKeyID),
SSEAlgorithm: c.String(mimirSSEAlgorithm),
}
cl, err := mimir.NewClient(ctx, mCfg)
if err != nil {
return fmt.Errorf("failed to create mimir client: %s", err)
}
pp := prometheusProcessor{
cl: cl,
im: importer,
cc: c.Int(mimirConcurrency),
isVerbose: c.Bool(globalVerbose),
}
return pp.run(ctx)
},
},
{
Name: "thanos",
Usage: "Migrate time series from Thanos blocks (supports raw and downsampled data)",
Flags: mergeFlags(globalFlags, thanosFlags, vmFlags),
Before: beforeFn,
Action: func(c *cli.Context) error {
fmt.Println("Thanos import mode")
vmCfg, err := initConfigVM(c)
if err != nil {
return fmt.Errorf("failed to init VM configuration: %s", err)
}
importer, err = vm.NewImporter(ctx, vmCfg)
if err != nil {
return fmt.Errorf("failed to create VM importer: %s", err)
}
thanosCfg := thanos.Config{
Snapshot: c.String(thanosSnapshot),
Filter: thanos.Filter{

View File

@@ -0,0 +1,195 @@
package mimir
import (
"fmt"
"log"
"os"
"path/filepath"
"sync"
"github.com/oklog/ulid/v2"
"github.com/prometheus/prometheus/tsdb"
"github.com/prometheus/prometheus/tsdb/tombstones"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/common"
)
var _ tsdb.BlockReader = (*lazyBlockReader)(nil)
// lazyBlockReader is stores block id and segment num information.
// It is used to lazily fetch and parse block data.
// It implements tsdb.BlockReader interface.
type lazyBlockReader struct {
// Block ID.
ID ulid.ULID
// SegmentsNum stores the number of chunks segments in the block.
SegmentsNum int
mu sync.Mutex
reader *tsdb.Block
tempDirPath string
fs common.RemoteFS
err error
}
// newLazyBlockReader returns a new LazyBlockReader for the given block.
func newLazyBlockReader(block *Block, fs common.RemoteFS) (*lazyBlockReader, error) {
if block.SegmentsFormat != "1b6d" {
return nil, fmt.Errorf("unsupported segments format: %s", block.SegmentsFormat)
}
return &lazyBlockReader{
ID: block.ID,
SegmentsNum: block.SegmentsNum,
fs: fs,
}, nil
}
func (lbr *lazyBlockReader) initialize() error {
lbr.mu.Lock()
defer lbr.mu.Unlock()
if lbr.reader != nil {
return nil
}
// fetching block and parse it and store it in lbr.reader
temp, err := lbr.mkTempDir()
if err != nil {
return fmt.Errorf("failed to create temp dir: %s", err)
}
lbr.tempDirPath = temp
// TODO: replace fetchFile and writeFile with buffered IO if needed
meta, err := lbr.fetchFile(metaFilename)
if err != nil {
return err
}
if err := lbr.writeFile(temp, metaFilename, meta); err != nil {
return fmt.Errorf("failed to write meta file: %w", err)
}
idx, err := lbr.fetchFile(indexFilename)
if err != nil {
return fmt.Errorf("failed to fetch index file %q: %w", indexFilename, err)
}
if err := lbr.writeFile(temp, indexFilename, idx); err != nil {
return err
}
for i := 1; i <= lbr.SegmentsNum; i++ {
// segments formats has format 1b06d
// https://github.com/grafana/mimir/blob/main/pkg/storage/tsdb/bucketindex/index.go#L32
chunkName := fmt.Sprintf("%06d", i)
blockChunkPath := filepath.Join("chunks", chunkName)
chunk, err := lbr.fetchFile(blockChunkPath)
if err != nil {
return fmt.Errorf("failed to fetch chunk file: %q: %w", chunkName, err)
}
if err := lbr.writeFile(temp, blockChunkPath, chunk); err != nil {
return fmt.Errorf("failed to write chunk file: %q: %s", chunkName, err)
}
}
// Set postingDecoder to nil because
// If it is nil then a default decoder is used, compatible with Prometheus v2.
pb, err := tsdb.OpenBlock(nil, temp, nil, nil)
if err != nil {
return fmt.Errorf("failed to open block %q: %w", lbr.ID, err)
}
lbr.reader = pb
return nil
}
// Index returns an IndexReader over the block's data.
func (lbr *lazyBlockReader) Index() (tsdb.IndexReader, error) {
if err := lbr.initialize(); err != nil {
return nil, err
}
return lbr.reader.Index()
}
// Chunks returns a ChunkReader over the block's data.
func (lbr *lazyBlockReader) Chunks() (tsdb.ChunkReader, error) {
if err := lbr.initialize(); err != nil {
return nil, err
}
return lbr.reader.Chunks()
}
// Tombstones returns a tombstones.Reader over the block's deleted data.
func (lbr *lazyBlockReader) Tombstones() (tombstones.Reader, error) {
if err := lbr.initialize(); err != nil {
return nil, err
}
return lbr.reader.Tombstones()
}
// Meta provides meta information about the block reader.
func (lbr *lazyBlockReader) Meta() tsdb.BlockMeta {
if err := lbr.initialize(); err != nil {
lbr.err = fmt.Errorf("cannot get BlockMeta: %w", err)
return tsdb.BlockMeta{}
}
return lbr.reader.Meta()
}
// Size returns the number of bytes that the block takes up on disk.
func (lbr *lazyBlockReader) Size() int64 {
if err := lbr.initialize(); err != nil {
lbr.err = fmt.Errorf("error get Size of the block: %s, return zero size", err)
return 0
}
return lbr.reader.Size()
}
// Err returns the last error that occurred on the block reader.
func (lbr *lazyBlockReader) Err() error {
return lbr.err
}
// Close closes block and releases all resources
func (lbr *lazyBlockReader) Close() error {
lbr.mu.Lock()
defer lbr.mu.Unlock()
if lbr.reader == nil {
return nil
}
err := lbr.reader.Close()
if err := os.RemoveAll(lbr.tempDirPath); err != nil {
log.Printf("failed to remove temp dir: %s", err)
}
lbr.reader = nil
lbr.tempDirPath = ""
return err
}
func (lbr *lazyBlockReader) mkTempDir() (string, error) {
temp, err := os.MkdirTemp("", lbr.ID.String())
if err != nil {
return "", fmt.Errorf("failed to create temp dir: %s", err)
}
err = os.Mkdir(filepath.Join(temp, "chunks"), os.ModePerm)
if err != nil {
return "", fmt.Errorf("failed to create temp dir: %s", err)
}
return temp, nil
}
func (lbr *lazyBlockReader) fetchFile(filePath string) ([]byte, error) {
blockID := lbr.ID.String()
blockPath := filepath.Join(blockID, filePath)
has, err := lbr.fs.HasFile(blockPath)
if err != nil {
return nil, err
}
if !has {
return nil, fmt.Errorf("block meta %s not found", blockID)
}
return lbr.fs.ReadFile(blockPath)
}
func (lbr *lazyBlockReader) writeFile(folder string, filename string, file []byte) error {
fileName := filepath.Join(folder, filename)
return os.WriteFile(fileName, file, os.ModePerm)
}

238
app/vmctl/mimir/mimir.go Normal file
View File

@@ -0,0 +1,238 @@
package mimir
import (
"bytes"
"compress/gzip"
"context"
"encoding/json"
"fmt"
"log"
"github.com/oklog/ulid/v2"
"github.com/prometheus/prometheus/model/labels"
"github.com/prometheus/prometheus/tsdb"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/prometheus"
utils "github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/vmctlutil"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/common"
)
const (
bucketIndex = "bucket-index.json"
bucketIndexCompressedFilename = bucketIndex + ".gz"
metaFilename = "meta.json"
indexFilename = "index"
)
// BlockDeletionMark holds the information about a block's deletion mark in the index.
// This type was copied from the mimir repository https://github.com/grafana/mimir/blob/main/pkg/storage/tsdb/bucketindex/index.go#L234.
type BlockDeletionMark struct {
// Block ID.
ID ulid.ULID `json:"block_id"`
// DeletionTime is a unix timestamp (seconds precision) of when the block was marked to be deleted.
DeletionTime int64 `json:"deletion_time"`
}
// Block holds the information about a block in the index.
// This is a partial implementation of the https://github.com/grafana/mimir/blob/main/pkg/storage/tsdb/bucketindex/index.go#L73
type Block struct {
// Block ID.
ID ulid.ULID `json:"block_id"`
// MinTime and MaxTime specify the time range all samples in the block are in (millis precision).
MinTime int64 `json:"min_time"`
MaxTime int64 `json:"max_time"`
// SegmentsFormat and SegmentsNum stores the format and number of chunks segments
// in the block.
SegmentsFormat string `json:"segments_format,omitempty"`
SegmentsNum int `json:"segments_num,omitempty"`
}
// Index contains all known blocks and markers of a tenant.
// This is a partial implementation pof the https://github.com/grafana/mimir/blob/main/pkg/storage/tsdb/bucketindex/index.go#L36
type Index struct {
// Version of the index format.
Version int `json:"version"`
// List of complete blocks (partial blocks are excluded from the index).
Blocks []*Block `json:"blocks"`
}
// Config contains a list of params needed
// for reading mimir snapshots
type Config struct {
// Path to remote storage bucket
Path string
// TenantID is the tenant id for the storage
TenantID string
Filter Filter
CredsFilePath string
ConfigFilePath string
ConfigProfile string
CustomS3Endpoint string
S3ForcePathStyle bool
S3TLSInsecureSkipVerify bool
SSEKMSKeyID string
SSEAlgorithm string
}
// Filter contains configuration for filtering
// the timeseries
type Filter struct {
TimeMin string
TimeMax string
Label string
LabelValue string
}
// Client is a wrapper over Prometheus tsdb.DBReader
type Client struct {
common.RemoteFS
filter filter
}
type filter struct {
min, max int64
label string
labelValue string
}
func (f filter) inRange(minTime, maxTime int64) bool {
fmin, fmax := f.min, f.max
if minTime == 0 {
fmin = minTime
}
if fmax == 0 {
fmax = maxTime
}
return minTime <= fmax && fmin <= maxTime
}
// NewClient creates and validates new Client
// with given Config
func NewClient(ctx context.Context, cfg Config) (*Client, error) {
if cfg.Path == "" {
return nil, fmt.Errorf("path cannot be empty")
}
if cfg.TenantID != "" {
cfg.Path = fmt.Sprintf("%s/%s", cfg.Path, cfg.TenantID)
}
var c Client
rfs, err := newRemoteFS(ctx, cfg)
if err != nil {
return nil, fmt.Errorf("cannot parse `-src`=%q: %w", cfg.Path, err)
}
c.RemoteFS = rfs
timeMin, err := utils.ParseTime(cfg.Filter.TimeMin)
if err != nil {
return nil, fmt.Errorf("failed to parse min time in filter: %s", err)
}
timeMax, err := utils.ParseTime(cfg.Filter.TimeMax)
if err != nil {
return nil, fmt.Errorf("failed to parse max time in filter: %s", err)
}
c.filter = filter{
min: timeMin.UnixMilli(),
max: timeMax.UnixMilli(),
label: cfg.Filter.Label,
labelValue: cfg.Filter.LabelValue,
}
return &c, nil
}
// Explore a fetches bucket-index.json file from a remote storage or local filesystem
// and filter blocks via the defined time range, but does not take into account label filters.
func (c *Client) Explore() ([]tsdb.BlockReader, error) {
log.Printf("Fetching blocks from remote storage")
indexFile, err := c.fetchIndexFile()
if err != nil {
return nil, fmt.Errorf("failed to fetch index file: %s", err)
}
var blocksToImport []tsdb.BlockReader
for _, block := range indexFile.Blocks {
if !c.filter.inRange(block.MinTime, block.MaxTime) {
// Skipping block outside of time range
continue
}
if block.ID.String() == "" {
continue
}
lazyBlockReader, err := newLazyBlockReader(block, c.RemoteFS)
if err != nil {
return nil, fmt.Errorf("failed to create lazy block reader: %s", err)
}
blocksToImport = append(blocksToImport, lazyBlockReader)
}
return blocksToImport, nil
}
// Read reads the given BlockReader according to configured
// time and label filters.
func (c *Client) Read(ctx context.Context, block tsdb.BlockReader) (*prometheus.CloseableSeriesSet, error) {
meta := block.Meta()
if b, ok := block.(*lazyBlockReader); ok && b.Err() != nil {
return nil, fmt.Errorf("failed to read block: %s", b.Err())
}
if meta.ULID.String() == "" {
return nil, fmt.Errorf("unexpected block without id")
}
minTime, maxTime := meta.MinTime, meta.MaxTime
if c.filter.min != 0 {
minTime = c.filter.min
}
if c.filter.max != 0 {
maxTime = c.filter.max
}
q, err := tsdb.NewBlockQuerier(block, minTime, maxTime)
if err != nil {
return nil, err
}
ss := q.Select(ctx, false, nil, labels.MustNewMatcher(labels.MatchRegexp, c.filter.label, c.filter.labelValue))
return &prometheus.CloseableSeriesSet{SeriesSet: ss, Close: q.Close}, nil
}
func (c *Client) fetchIndexFile() (*Index, error) {
has, err := c.HasFile(bucketIndexCompressedFilename)
if err != nil {
return nil, err
}
if !has {
return nil, fmt.Errorf("bucket-index.json.gz not found")
}
file, err := c.ReadFile(bucketIndexCompressedFilename)
if err != nil {
return nil, fmt.Errorf("failed to read bucket index: %s", err)
}
r := bytes.NewReader(file)
// Read all the content.
gzipReader, err := gzip.NewReader(r)
if err != nil {
return nil, fmt.Errorf("failed to create gzip reader: %s", err)
}
var indexFile Index
err = json.NewDecoder(gzipReader).Decode(&indexFile)
if err != nil {
return nil, fmt.Errorf("failed to decode bucket index: %s", err)
}
return &indexFile, nil
}

View File

@@ -0,0 +1,93 @@
package mimir
import (
"context"
"fmt"
"path/filepath"
"strings"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/azremote"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/common"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/fsremote"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/gcsremote"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/backup/s3remote"
)
// newRemoteFS returns new remote fs from the given Config.
func newRemoteFS(ctx context.Context, cfg Config) (common.RemoteFS, error) {
if len(cfg.Path) == 0 {
return nil, fmt.Errorf("path cannot be empty")
}
n := strings.Index(cfg.Path, "://")
if n < 0 {
return nil, fmt.Errorf("missing scheme in path %q. Supported schemes: `gs://`, `s3://`, `azblob://`, `fs://`", cfg.Path)
}
scheme := cfg.Path[:n]
dir := cfg.Path[n+len("://"):]
switch scheme {
case "fs":
if !filepath.IsAbs(dir) {
return nil, fmt.Errorf("dir must be absolute; got %q", dir)
}
fsr := &fsremote.FS{
Dir: filepath.Clean(dir),
}
return fsr, nil
case "gcs", "gs":
n := strings.Index(dir, "/")
if n < 0 {
return nil, fmt.Errorf("missing directory on the gcs bucket %q", dir)
}
bucket := dir[:n]
dir = dir[n:]
fsr := &gcsremote.FS{
CredsFilePath: cfg.CredsFilePath,
Bucket: bucket,
Dir: dir,
}
if err := fsr.Init(ctx); err != nil {
return nil, fmt.Errorf("cannot initialize connection to gcs: %w", err)
}
return fsr, nil
case "azblob":
n := strings.Index(dir, "/")
if n < 0 {
return nil, fmt.Errorf("missing directory on the AZBlob container %q", dir)
}
bucket := dir[:n]
dir = dir[n:]
fsr := &azremote.FS{
Container: bucket,
Dir: dir,
}
if err := fsr.Init(ctx); err != nil {
return nil, fmt.Errorf("cannot initialize connection to AZBlob: %w", err)
}
return fsr, nil
case "s3":
n := strings.Index(dir, "/")
if n < 0 {
return nil, fmt.Errorf("missing directory on the s3 bucket %q", dir)
}
bucket := dir[:n]
dir = dir[n:]
fsr := &s3remote.FS{
CredsFilePath: cfg.CredsFilePath,
ConfigFilePath: cfg.ConfigFilePath,
CustomEndpoint: cfg.CustomS3Endpoint,
TLSInsecureSkipVerify: cfg.S3TLSInsecureSkipVerify,
S3ForcePathStyle: cfg.S3ForcePathStyle,
ProfileName: cfg.ConfigProfile,
Bucket: bucket,
Dir: dir,
SSEKMSKeyId: cfg.SSEKMSKeyID,
SSEAlgorithm: s3remote.StringToEncryptionAlgorithm(cfg.SSEAlgorithm),
}
if err := fsr.Init(ctx); err != nil {
return nil, fmt.Errorf("cannot initialize connection to s3: %w", err)
}
return fsr, nil
default:
return nil, fmt.Errorf("unsupported scheme %q", scheme)
}
}

View File

@@ -3,6 +3,7 @@ package main
import (
"context"
"fmt"
"io"
"log"
"strings"
"sync"
@@ -18,10 +19,17 @@ import (
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/vm"
)
// Runner is an interface for fetching and reading
// snapshot blocks
type Runner interface {
Explore() ([]tsdb.BlockReader, error)
Read(context.Context, tsdb.BlockReader) (*prometheus.CloseableSeriesSet, error)
}
type prometheusProcessor struct {
// prometheus client fetches and reads
// Runner fetches and reads
// snapshot blocks
cl *prometheus.Client
cl Runner
// importer performs import requests
// for timeseries data returned from
// snapshot blocks
@@ -48,7 +56,7 @@ func (pp *prometheusProcessor) run(ctx context.Context) error {
return nil
}
if err := pp.processBlocks(blocks); err != nil {
if err := pp.processBlocks(ctx, blocks); err != nil {
return fmt.Errorf("migration failed: %s", err)
}
@@ -57,11 +65,17 @@ func (pp *prometheusProcessor) run(ctx context.Context) error {
return nil
}
func (pp *prometheusProcessor) do(b tsdb.BlockReader) error {
ss, err := pp.cl.Read(b)
func (pp *prometheusProcessor) do(ctx context.Context, b tsdb.BlockReader) error {
css, err := pp.cl.Read(ctx, b)
if err != nil {
return fmt.Errorf("failed to read block: %s", err)
}
defer func() {
if err := css.Close(); err != nil {
log.Printf("cannot close SeriesSet for block: %q : %s\n", b.Meta().ULID, err)
}
}()
ss := css.SeriesSet
var it chunkenc.Iterator
for ss.Next() {
var name string
@@ -114,7 +128,7 @@ func (pp *prometheusProcessor) do(b tsdb.BlockReader) error {
return ss.Err()
}
func (pp *prometheusProcessor) processBlocks(blocks []tsdb.BlockReader) error {
func (pp *prometheusProcessor) processBlocks(ctx context.Context, blocks []tsdb.BlockReader) error {
promBlocksTotal.Add(len(blocks))
bar := barpool.AddWithTemplate(fmt.Sprintf(barTpl, "Processing blocks"), len(blocks))
if err := barpool.Start(); err != nil {
@@ -130,11 +144,16 @@ func (pp *prometheusProcessor) processBlocks(blocks []tsdb.BlockReader) error {
for range pp.cc {
wg.Go(func() {
for br := range blockReadersCh {
if err := pp.do(br); err != nil {
if err := pp.do(ctx, br); err != nil {
promErrorsTotal.Inc()
errCh <- fmt.Errorf("read failed for block %q: %s", br.Meta().ULID, err)
errCh <- fmt.Errorf("cannot read block %q: %s", br.Meta().ULID, err)
return
}
if cb, ok := br.(io.Closer); ok {
if err := cb.Close(); err != nil {
errCh <- fmt.Errorf("cannot close block: %q: %w", br.Meta().ULID, err)
}
}
promBlocksProcessed.Inc()
bar.Increment()
}

View File

@@ -8,6 +8,8 @@ import (
"github.com/prometheus/prometheus/model/labels"
"github.com/prometheus/prometheus/storage"
"github.com/prometheus/prometheus/tsdb"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmctl/vmctlutil"
)
// Config contains a list of params needed
@@ -60,13 +62,13 @@ func NewClient(cfg Config) (*Client, error) {
return nil, fmt.Errorf("failed to open snapshot %q: %s", cfg.Snapshot, err)
}
c := &Client{DBReadOnly: db}
minTime, maxTime, err := parseTime(cfg.Filter.TimeMin, cfg.Filter.TimeMax)
timeMin, timeMax, err := parseTime(cfg.Filter.TimeMin, cfg.Filter.TimeMax)
if err != nil {
return nil, fmt.Errorf("failed to parse time in filter: %s", err)
}
c.filter = filter{
min: minTime,
max: maxTime,
min: timeMin,
max: timeMax,
label: cfg.Filter.Label,
labelValue: cfg.Filter.LabelValue,
}
@@ -83,7 +85,7 @@ func (c *Client) Explore() ([]tsdb.BlockReader, error) {
if err != nil {
return nil, fmt.Errorf("failed to fetch blocks: %s", err)
}
s := &Stats{
s := &vmctlutil.Stats{
Filtered: c.filter.min != 0 || c.filter.max != 0 || c.filter.label != "",
Blocks: len(blocks),
}
@@ -108,9 +110,15 @@ func (c *Client) Explore() ([]tsdb.BlockReader, error) {
return blocksToImport, nil
}
// CloseableSeriesSet defines a SeriesSet with Close method
type CloseableSeriesSet struct {
SeriesSet storage.SeriesSet
Close func() error
}
// Read reads the given BlockReader according to configured
// time and label filters.
func (c *Client) Read(block tsdb.BlockReader) (storage.SeriesSet, error) {
func (c *Client) Read(ctx context.Context, block tsdb.BlockReader) (*CloseableSeriesSet, error) {
minTime, maxTime := block.Meta().MinTime, block.Meta().MaxTime
if c.filter.min != 0 {
minTime = c.filter.min
@@ -122,8 +130,8 @@ func (c *Client) Read(block tsdb.BlockReader) (storage.SeriesSet, error) {
if err != nil {
return nil, err
}
ss := q.Select(context.Background(), false, nil, labels.MustNewMatcher(labels.MatchRegexp, c.filter.label, c.filter.labelValue))
return ss, nil
ss := q.Select(ctx, false, nil, labels.MustNewMatcher(labels.MatchRegexp, c.filter.label, c.filter.labelValue))
return &CloseableSeriesSet{ss, q.Close}, nil
}
func parseTime(start, end string) (int64, int64, error) {

View File

@@ -1,4 +1,4 @@
package prometheus
package vmctlutil
import (
"fmt"
@@ -18,7 +18,7 @@ type Stats struct {
// String returns string representation for s.
func (s Stats) String() string {
str := fmt.Sprintf("Prometheus snapshot stats:\n"+
str := fmt.Sprintf("Snapshot stats:\n"+
" blocks found: %d;\n"+
" blocks skipped by time filter: %d;\n"+
" min time: %d (%v);\n"+

View File

@@ -2,13 +2,16 @@
"math"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmselect/netstorage"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/bytesutil"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/promrelabel"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/storage"
) %}
{% stripspace %}
// Federate writes rs in /federate format.
// See https://prometheus.io/docs/prometheus/latest/federation/
{% func Federate(rs *netstorage.Result) %}
{% func Federate(rs *netstorage.Result, escapeScheme string) %}
{% code
values := rs.Values
timestamps := rs.Timestamps
@@ -24,10 +27,54 @@
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3185
{% endcomment %}
{% return %}
{% endif %}
{%= prometheusMetricName(&rs.MetricName) %}{% space %}
{% endif %}
{% switch escapeScheme %}
{% case federateEscapeSchemeUTF8 %}
{%= prometheusFederateMetricNameUTF8(&rs.MetricName) %}{% space %}
{% case federateEscapeSchemeUnderscore %}
{%= prometheusFederateMetricNameEscapeUnderscore(&rs.MetricName) %}{% space %}
{% case "" %}
{%= prometheusMetricName(&rs.MetricName) %}{% space %}
{% endswitch %}
{%f= lastValue %}{% space %}
{%dl= timestamps[len(timestamps)-1] %}{% newline %}
{% endfunc %}
{% func prometheusFederateMetricNameEscapeUnderscore(mn *storage.MetricName) %}
{%s= promrelabel.SanitizeMetricName(bytesutil.ToUnsafeString(mn.MetricGroup)) %}
{% if len(mn.Tags) > 0 %}
{
{% code tags := mn.Tags %}
{%s= promrelabel.SanitizeLabelName(bytesutil.ToUnsafeString(tags[0].Key)) %}={%= escapePrometheusLabel(tags[0].Value) %}
{% code tags = tags[1:] %}
{% for i := range tags %}
{% code tag := &tags[i] %}
,{%s= promrelabel.SanitizeLabelName(bytesutil.ToUnsafeString(tag.Key)) %}={%= escapePrometheusLabel(tag.Value) %}
{% endfor %}
}
{% endif %}
{% endfunc %}
{% func prometheusFederateMetricNameUTF8(mn *storage.MetricName) %}
{
{%= escapePrometheusLabel(mn.MetricGroup) %}
{% if len(mn.Tags) > 0 %}
,
{% code tags := mn.Tags %}
{%= escapePrometheusLabel(tags[0].Key) %}={%= escapePrometheusLabel(tags[0].Value) %}
{% code tags = tags[1:] %}
{% for i := range tags %}
{% code tag := &tags[i] %}
,{%= escapePrometheusLabel(tag.Key) %}={%= escapePrometheusLabel(tag.Value) %}
{% endfor %}
{% endif %}
}
{% endfunc %}
{% endstripspace %}

View File

@@ -9,82 +9,241 @@ import (
"math"
"github.com/VictoriaMetrics/VictoriaMetrics/app/vmselect/netstorage"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/bytesutil"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/promrelabel"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/storage"
)
// Federate writes rs in /federate format.// See https://prometheus.io/docs/prometheus/latest/federation/
//line app/vmselect/prometheus/federate.qtpl:11
//line app/vmselect/prometheus/federate.qtpl:14
import (
qtio422016 "io"
qt422016 "github.com/valyala/quicktemplate"
)
//line app/vmselect/prometheus/federate.qtpl:11
//line app/vmselect/prometheus/federate.qtpl:14
var (
_ = qtio422016.Copy
_ = qt422016.AcquireByteBuffer
)
//line app/vmselect/prometheus/federate.qtpl:11
func StreamFederate(qw422016 *qt422016.Writer, rs *netstorage.Result) {
//line app/vmselect/prometheus/federate.qtpl:13
//line app/vmselect/prometheus/federate.qtpl:14
func StreamFederate(qw422016 *qt422016.Writer, rs *netstorage.Result, escapeScheme string) {
//line app/vmselect/prometheus/federate.qtpl:16
values := rs.Values
timestamps := rs.Timestamps
//line app/vmselect/prometheus/federate.qtpl:16
//line app/vmselect/prometheus/federate.qtpl:19
if len(timestamps) == 0 || len(values) == 0 {
//line app/vmselect/prometheus/federate.qtpl:16
//line app/vmselect/prometheus/federate.qtpl:19
return
//line app/vmselect/prometheus/federate.qtpl:16
//line app/vmselect/prometheus/federate.qtpl:19
}
//line app/vmselect/prometheus/federate.qtpl:18
//line app/vmselect/prometheus/federate.qtpl:21
lastValue := values[len(values)-1]
//line app/vmselect/prometheus/federate.qtpl:20
//line app/vmselect/prometheus/federate.qtpl:23
if math.IsNaN(lastValue) {
//line app/vmselect/prometheus/federate.qtpl:26
//line app/vmselect/prometheus/federate.qtpl:29
return
//line app/vmselect/prometheus/federate.qtpl:27
//line app/vmselect/prometheus/federate.qtpl:30
}
//line app/vmselect/prometheus/federate.qtpl:28
streamprometheusMetricName(qw422016, &rs.MetricName)
//line app/vmselect/prometheus/federate.qtpl:28
qw422016.N().S(` `)
//line app/vmselect/prometheus/federate.qtpl:29
//line app/vmselect/prometheus/federate.qtpl:32
switch escapeScheme {
//line app/vmselect/prometheus/federate.qtpl:33
case federateEscapeSchemeUTF8:
//line app/vmselect/prometheus/federate.qtpl:34
streamprometheusFederateMetricNameUTF8(qw422016, &rs.MetricName)
//line app/vmselect/prometheus/federate.qtpl:34
qw422016.N().S(` `)
//line app/vmselect/prometheus/federate.qtpl:36
case federateEscapeSchemeUnderscore:
//line app/vmselect/prometheus/federate.qtpl:37
streamprometheusFederateMetricNameEscapeUnderscore(qw422016, &rs.MetricName)
//line app/vmselect/prometheus/federate.qtpl:37
qw422016.N().S(` `)
//line app/vmselect/prometheus/federate.qtpl:39
case "":
//line app/vmselect/prometheus/federate.qtpl:40
streamprometheusMetricName(qw422016, &rs.MetricName)
//line app/vmselect/prometheus/federate.qtpl:40
qw422016.N().S(` `)
//line app/vmselect/prometheus/federate.qtpl:41
}
//line app/vmselect/prometheus/federate.qtpl:43
qw422016.N().F(lastValue)
//line app/vmselect/prometheus/federate.qtpl:29
//line app/vmselect/prometheus/federate.qtpl:43
qw422016.N().S(` `)
//line app/vmselect/prometheus/federate.qtpl:30
//line app/vmselect/prometheus/federate.qtpl:44
qw422016.N().DL(timestamps[len(timestamps)-1])
//line app/vmselect/prometheus/federate.qtpl:30
//line app/vmselect/prometheus/federate.qtpl:44
qw422016.N().S(`
`)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
}
//line app/vmselect/prometheus/federate.qtpl:31
func WriteFederate(qq422016 qtio422016.Writer, rs *netstorage.Result) {
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
func WriteFederate(qq422016 qtio422016.Writer, rs *netstorage.Result, escapeScheme string) {
//line app/vmselect/prometheus/federate.qtpl:45
qw422016 := qt422016.AcquireWriter(qq422016)
//line app/vmselect/prometheus/federate.qtpl:31
StreamFederate(qw422016, rs)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
StreamFederate(qw422016, rs, escapeScheme)
//line app/vmselect/prometheus/federate.qtpl:45
qt422016.ReleaseWriter(qw422016)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
}
//line app/vmselect/prometheus/federate.qtpl:31
func Federate(rs *netstorage.Result) string {
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
func Federate(rs *netstorage.Result, escapeScheme string) string {
//line app/vmselect/prometheus/federate.qtpl:45
qb422016 := qt422016.AcquireByteBuffer()
//line app/vmselect/prometheus/federate.qtpl:31
WriteFederate(qb422016, rs)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
WriteFederate(qb422016, rs, escapeScheme)
//line app/vmselect/prometheus/federate.qtpl:45
qs422016 := string(qb422016.B)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
qt422016.ReleaseByteBuffer(qb422016)
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
return qs422016
//line app/vmselect/prometheus/federate.qtpl:31
//line app/vmselect/prometheus/federate.qtpl:45
}
//line app/vmselect/prometheus/federate.qtpl:47
func streamprometheusFederateMetricNameEscapeUnderscore(qw422016 *qt422016.Writer, mn *storage.MetricName) {
//line app/vmselect/prometheus/federate.qtpl:48
qw422016.N().S(promrelabel.SanitizeMetricName(bytesutil.ToUnsafeString(mn.MetricGroup)))
//line app/vmselect/prometheus/federate.qtpl:49
if len(mn.Tags) > 0 {
//line app/vmselect/prometheus/federate.qtpl:49
qw422016.N().S(`{`)
//line app/vmselect/prometheus/federate.qtpl:51
tags := mn.Tags
//line app/vmselect/prometheus/federate.qtpl:52
qw422016.N().S(promrelabel.SanitizeLabelName(bytesutil.ToUnsafeString(tags[0].Key)))
//line app/vmselect/prometheus/federate.qtpl:52
qw422016.N().S(`=`)
//line app/vmselect/prometheus/federate.qtpl:52
streamescapePrometheusLabel(qw422016, tags[0].Value)
//line app/vmselect/prometheus/federate.qtpl:53
tags = tags[1:]
//line app/vmselect/prometheus/federate.qtpl:54
for i := range tags {
//line app/vmselect/prometheus/federate.qtpl:55
tag := &tags[i]
//line app/vmselect/prometheus/federate.qtpl:55
qw422016.N().S(`,`)
//line app/vmselect/prometheus/federate.qtpl:56
qw422016.N().S(promrelabel.SanitizeLabelName(bytesutil.ToUnsafeString(tag.Key)))
//line app/vmselect/prometheus/federate.qtpl:56
qw422016.N().S(`=`)
//line app/vmselect/prometheus/federate.qtpl:56
streamescapePrometheusLabel(qw422016, tag.Value)
//line app/vmselect/prometheus/federate.qtpl:57
}
//line app/vmselect/prometheus/federate.qtpl:57
qw422016.N().S(`}`)
//line app/vmselect/prometheus/federate.qtpl:59
}
//line app/vmselect/prometheus/federate.qtpl:60
}
//line app/vmselect/prometheus/federate.qtpl:60
func writeprometheusFederateMetricNameEscapeUnderscore(qq422016 qtio422016.Writer, mn *storage.MetricName) {
//line app/vmselect/prometheus/federate.qtpl:60
qw422016 := qt422016.AcquireWriter(qq422016)
//line app/vmselect/prometheus/federate.qtpl:60
streamprometheusFederateMetricNameEscapeUnderscore(qw422016, mn)
//line app/vmselect/prometheus/federate.qtpl:60
qt422016.ReleaseWriter(qw422016)
//line app/vmselect/prometheus/federate.qtpl:60
}
//line app/vmselect/prometheus/federate.qtpl:60
func prometheusFederateMetricNameEscapeUnderscore(mn *storage.MetricName) string {
//line app/vmselect/prometheus/federate.qtpl:60
qb422016 := qt422016.AcquireByteBuffer()
//line app/vmselect/prometheus/federate.qtpl:60
writeprometheusFederateMetricNameEscapeUnderscore(qb422016, mn)
//line app/vmselect/prometheus/federate.qtpl:60
qs422016 := string(qb422016.B)
//line app/vmselect/prometheus/federate.qtpl:60
qt422016.ReleaseByteBuffer(qb422016)
//line app/vmselect/prometheus/federate.qtpl:60
return qs422016
//line app/vmselect/prometheus/federate.qtpl:60
}
//line app/vmselect/prometheus/federate.qtpl:62
func streamprometheusFederateMetricNameUTF8(qw422016 *qt422016.Writer, mn *storage.MetricName) {
//line app/vmselect/prometheus/federate.qtpl:62
qw422016.N().S(`{`)
//line app/vmselect/prometheus/federate.qtpl:64
streamescapePrometheusLabel(qw422016, mn.MetricGroup)
//line app/vmselect/prometheus/federate.qtpl:65
if len(mn.Tags) > 0 {
//line app/vmselect/prometheus/federate.qtpl:65
qw422016.N().S(`,`)
//line app/vmselect/prometheus/federate.qtpl:67
tags := mn.Tags
//line app/vmselect/prometheus/federate.qtpl:68
streamescapePrometheusLabel(qw422016, tags[0].Key)
//line app/vmselect/prometheus/federate.qtpl:68
qw422016.N().S(`=`)
//line app/vmselect/prometheus/federate.qtpl:68
streamescapePrometheusLabel(qw422016, tags[0].Value)
//line app/vmselect/prometheus/federate.qtpl:69
tags = tags[1:]
//line app/vmselect/prometheus/federate.qtpl:70
for i := range tags {
//line app/vmselect/prometheus/federate.qtpl:71
tag := &tags[i]
//line app/vmselect/prometheus/federate.qtpl:71
qw422016.N().S(`,`)
//line app/vmselect/prometheus/federate.qtpl:72
streamescapePrometheusLabel(qw422016, tag.Key)
//line app/vmselect/prometheus/federate.qtpl:72
qw422016.N().S(`=`)
//line app/vmselect/prometheus/federate.qtpl:72
streamescapePrometheusLabel(qw422016, tag.Value)
//line app/vmselect/prometheus/federate.qtpl:73
}
//line app/vmselect/prometheus/federate.qtpl:74
}
//line app/vmselect/prometheus/federate.qtpl:74
qw422016.N().S(`}`)
//line app/vmselect/prometheus/federate.qtpl:76
}
//line app/vmselect/prometheus/federate.qtpl:76
func writeprometheusFederateMetricNameUTF8(qq422016 qtio422016.Writer, mn *storage.MetricName) {
//line app/vmselect/prometheus/federate.qtpl:76
qw422016 := qt422016.AcquireWriter(qq422016)
//line app/vmselect/prometheus/federate.qtpl:76
streamprometheusFederateMetricNameUTF8(qw422016, mn)
//line app/vmselect/prometheus/federate.qtpl:76
qt422016.ReleaseWriter(qw422016)
//line app/vmselect/prometheus/federate.qtpl:76
}
//line app/vmselect/prometheus/federate.qtpl:76
func prometheusFederateMetricNameUTF8(mn *storage.MetricName) string {
//line app/vmselect/prometheus/federate.qtpl:76
qb422016 := qt422016.AcquireByteBuffer()
//line app/vmselect/prometheus/federate.qtpl:76
writeprometheusFederateMetricNameUTF8(qb422016, mn)
//line app/vmselect/prometheus/federate.qtpl:76
qs422016 := string(qb422016.B)
//line app/vmselect/prometheus/federate.qtpl:76
qt422016.ReleaseByteBuffer(qb422016)
//line app/vmselect/prometheus/federate.qtpl:76
return qs422016
//line app/vmselect/prometheus/federate.qtpl:76
}

View File

@@ -8,15 +8,15 @@ import (
)
func TestFederate(t *testing.T) {
f := func(rs *netstorage.Result, expectedResult string) {
f := func(rs *netstorage.Result, escapeScheme string, expectedResult string) {
t.Helper()
result := Federate(rs)
result := Federate(rs, escapeScheme)
if result != expectedResult {
t.Fatalf("unexpected result; got\n%s\nwant\n%s", result, expectedResult)
}
}
f(&netstorage.Result{}, ``)
f(&netstorage.Result{}, ``, ``)
f(&netstorage.Result{
MetricName: storage.MetricName{
@@ -39,5 +39,60 @@ func TestFederate(t *testing.T) {
},
Values: []float64{1.23},
Timestamps: []int64{123},
}, `foo{a="b",qqq="\\",abc="a<b\"\\c"} 1.23 123`+"\n")
}, ``, `foo{a="b",qqq="\\",abc="a<b\"\\c"} 1.23 123`+"\n")
f(&netstorage.Result{
MetricName: storage.MetricName{
MetricGroup: []byte("foo.bar"),
Tags: []storage.Tag{
{
Key: []byte("some.!other"),
Value: []byte("value.unchanged!."),
},
{
Key: []byte("qqq"),
Value: []byte("\\"),
},
{
Key: []byte("!key"),
Value: []byte("value"),
},
{
Key: []byte("abc"),
// Verify that < isn't encoded. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/5431
Value: []byte("a<b\"\\c"),
},
},
},
Values: []float64{1.23},
Timestamps: []int64{123},
}, federateEscapeSchemeUnderscore, `foo_bar{some__other="value.unchanged!.",qqq="\\",_key="value",abc="a<b\"\\c"} 1.23 123`+"\n")
f(&netstorage.Result{
MetricName: storage.MetricName{
MetricGroup: []byte("foo.bar"),
Tags: []storage.Tag{
{
Key: []byte("some.!other"),
Value: []byte("value.unchanged!."),
},
{
Key: []byte("qqq"),
Value: []byte("\\"),
},
{
Key: []byte("!key"),
Value: []byte("value"),
},
{
Key: []byte(`ab"c`),
// Verify that < isn't encoded. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/5431
Value: []byte("a<b\"\\c"),
},
},
},
Values: []float64{1.23},
Timestamps: []int64{123},
}, federateEscapeSchemeUTF8, `{"foo.bar","some.!other"="value.unchanged!.","qqq"="\\","!key"="value","ab\"c"="a<b\"\\c"} 1.23 123`+"\n")
}

View File

@@ -9,16 +9,17 @@ import (
)
func BenchmarkFederate(b *testing.B) {
rs := &netstorage.Result{
MetricName: storage.MetricName{
MetricGroup: []byte("foo_bar_bazaaaa_total"),
MetricGroup: []byte("foo_bar_?_._bazaaaa_total"),
Tags: []storage.Tag{
{
Key: []byte("instance"),
Key: []byte("instance:job"),
Value: []byte("foobarbaz:2344"),
},
{
Key: []byte("job"),
Key: []byte("job.name"),
Value: []byte("aaabbbccc"),
},
},
@@ -27,12 +28,22 @@ func BenchmarkFederate(b *testing.B) {
Timestamps: []int64{1234567890},
}
b.ReportAllocs()
b.RunParallel(func(pb *testing.PB) {
var bb bytes.Buffer
for pb.Next() {
bb.Reset()
WriteFederate(&bb, rs)
}
})
f := func(name, escapeScheme string) {
b.Helper()
b.Run(name, func(b *testing.B) {
b.ReportAllocs()
b.RunParallel(func(pb *testing.PB) {
var bb bytes.Buffer
for pb.Next() {
bb.Reset()
WriteFederate(&bb, rs, escapeScheme)
}
})
})
}
f("without escape", "")
f("allow-utf-8", federateEscapeSchemeUTF8)
f("legacy-underscore", federateEscapeSchemeUnderscore)
}

View File

@@ -108,6 +108,11 @@ func PrettifyQuery(w http.ResponseWriter, r *http.Request) {
_ = bw.Flush()
}
const (
federateEscapeSchemeUnderscore = "underscore"
federateEscapeSchemeUTF8 = "utf-8"
)
// FederateHandler implements /federate . See https://prometheus.io/docs/prometheus/latest/federation/
func FederateHandler(startTime time.Time, w http.ResponseWriter, r *http.Request) error {
defer federateDuration.UpdateDuration(startTime)
@@ -132,6 +137,21 @@ func FederateHandler(startTime time.Time, w http.ResponseWriter, r *http.Request
return fmt.Errorf("cannot fetch data for %q: %w", sq, err)
}
// add best-effort format negotiation
// modern version of Prometheus always set allow-utf-8 in order to properly parse utf-8 names and labels
// prometheus below v3 uses underscore escaping by default and it's the most common standard
var escapeScheme string
accept := r.Header.Get("Accept")
if len(accept) > 0 && strings.Contains(accept, "allow-utf-8") {
escapeScheme = federateEscapeSchemeUTF8
}
// try fallback to legacy underscore escaping if needed for Prometheus only,
// it's not widely used after Prometheus v3.0 release
// most of the Prometheus scrapers already use allow-utf-8 header
isPrometheus := strings.HasPrefix(r.UserAgent(), "Prometheus")
if len(escapeScheme) == 0 && isPrometheus {
escapeScheme = federateEscapeSchemeUnderscore
}
w.Header().Set("Content-Type", "text/plain; charset=utf-8")
bw := bufferedwriter.Get(w)
defer bufferedwriter.Put(bw)
@@ -141,7 +161,7 @@ func FederateHandler(startTime time.Time, w http.ResponseWriter, r *http.Request
return err
}
bb := sw.getBuffer(workerID)
WriteFederate(bb, rs)
WriteFederate(bb, rs, escapeScheme)
return sw.maybeFlushBuffer(bb)
})
if err == nil {

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

View File

@@ -37,11 +37,11 @@
<meta property="og:title" content="UI for VictoriaMetrics">
<meta property="og:url" content="https://victoriametrics.com/">
<meta property="og:description" content="Explore and troubleshoot your VictoriaMetrics data">
<script type="module" crossorigin src="./assets/index-BjJ7fDL7.js"></script>
<script type="module" crossorigin src="./assets/index-CoGukb-x.js"></script>
<link rel="modulepreload" crossorigin href="./assets/rolldown-runtime-COnpUsM8.js">
<link rel="modulepreload" crossorigin href="./assets/vendor-C8Kwp93_.js">
<link rel="stylesheet" crossorigin href="./assets/vendor-CnsZ1jie.css">
<link rel="stylesheet" crossorigin href="./assets/index-BL7jEFBa.css">
<link rel="stylesheet" crossorigin href="./assets/index-BBUnmLOr.css">
</head>
<body>
<noscript>You need to enable JavaScript to run this app.</noscript>

View File

@@ -1,4 +1,4 @@
FROM golang:1.26.3 AS build-web-stage
FROM golang:1.26.4 AS build-web-stage
COPY build /build
WORKDIR /build

View File

@@ -1,7 +1,7 @@
export const seriesFetchedWarning = `No match!
export const seriesFetchedWarning = `No match!
This query hasn't selected any time series from database.
Either the requested metrics are missing in the database,
or there is a typo in series selector.`;
export const partialWarning = `The shown results are marked as PARTIAL.
The result is marked as partial if one or more vmstorage nodes failed to respond to the query.`;
The result is marked as partial if one or more storage nodes failed to respond to the query.`;

View File

@@ -71,7 +71,7 @@ const RulesHeader = ({
<TextField
label="Search"
value={search}
placeholder="Filter by rule, name or labels"
placeholder="Filter by group or rule name"
startIcon={<SearchIcon />}
onChange={onChangeSearch}
/>

View File

@@ -79,24 +79,25 @@ type PrometheusWriteQuerier interface {
// QueryOpts contains various params used for querying or ingesting data
type QueryOpts struct {
Tenant string
Timeout string
Start string
End string
Time string
Step string
ExtraFilters []string
ExtraLabels []string
Trace string
ReduceMemUsage string
MaxLookback string
LatencyOffset string
Format string
NoCache string
Headers http.Header
From string
Until string
StorageStep string
Tenant string
Timeout string
Start string
End string
Time string
Step string
ExtraFilters []string
ExtraLabels []string
Trace string
ReduceMemUsage string
MaxLookback string
LatencyOffset string
Format string
NoCache string
Headers http.Header
From string
Until string
StorageStep string
DenyPartialResponse string
}
func (qos *QueryOpts) getHeaders() http.Header {
@@ -132,6 +133,7 @@ func (qos *QueryOpts) asURLValues() url.Values {
addNonEmpty("from", qos.From)
addNonEmpty("until", qos.Until)
addNonEmpty("storage_step", qos.StorageStep)
addNonEmpty("deny_partial_response", qos.DenyPartialResponse)
return uv
}

View File

@@ -1015,35 +1015,42 @@ func testGroupSkipSlowReplicas(tc *apptest.TestCase, opts *testGroupReplicationO
func testGroupPartialResponse(tc *apptest.TestCase, opts *testGroupReplicationOpts) {
t := tc.T()
assertSeries := func(app *apptest.Vmselect, wantPartial bool) {
assertSeries := func(app *apptest.Vmselect, denyPartialResponse string, want *apptest.PrometheusAPIV1SeriesResponse) {
t.Helper()
tc.Assert(&apptest.AssertOptions{
Msg: "unexpected /api/v1/series response",
Got: func() any {
return app.PrometheusAPIV1Series(t, `{__name__=~".*"}`, apptest.QueryOpts{
Start: "2024-01-01T00:00:00Z",
End: "2024-01-31T00:00:00Z",
Start: "2024-01-01T00:00:00Z",
End: "2024-01-31T00:00:00Z",
DenyPartialResponse: denyPartialResponse,
}).Sort()
},
Want: &apptest.PrometheusAPIV1SeriesResponse{
Status: "success",
IsPartial: wantPartial,
},
Want: want,
CmpOpts: []cmp.Option{
cmpopts.IgnoreFields(apptest.PrometheusAPIV1SeriesResponse{}, "Data"),
cmpopts.IgnoreFields(apptest.PrometheusAPIV1SeriesResponse{}, "Data", "Error"),
},
})
}
mustReturnPartialResponse := true
mustReturnFullResponse := false
allowPartialResponse := ""
denyPartialResponse := "1"
mustReturnPartialResponse := &apptest.PrometheusAPIV1SeriesResponse{
Status: "success",
IsPartial: true,
}
mustReturnFullResponse := &apptest.PrometheusAPIV1SeriesResponse{
Status: "success",
IsPartial: false,
}
// All vmstorage replicas are available so both vmselects must return full
// response.
assertSeries(opts.c.vmselect, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnFullResponse)
// Stop groupRF-1 vmstorage nodes in first group.
//
@@ -1053,10 +1060,10 @@ func testGroupPartialResponse(tc *apptest.TestCase, opts *testGroupReplicationOp
// about the replication factor and therefore they must still be able to
// return full dataset.
opts.c.storageGroups[0].stopNodes(tc, opts.groupRF-1)
assertSeries(opts.c.vmselect, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnFullResponse)
// Stop groupRF-1 vmstorages in the remaining groups.
//
@@ -1066,10 +1073,10 @@ func testGroupPartialResponse(tc *apptest.TestCase, opts *testGroupReplicationOp
for g := 1; g < len(opts.c.storageGroups); g++ {
opts.c.storageGroups[g].stopNodes(tc, opts.groupRF-1)
}
assertSeries(opts.c.vmselect, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnFullResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnFullResponse)
// Stop one more vmstorage in the first group.
//
@@ -1077,10 +1084,10 @@ func testGroupPartialResponse(tc *apptest.TestCase, opts *testGroupReplicationOp
// because it is unaware of replication across groups. vmselectGroupGlobalRF
// will continue retuning full dataset.
opts.c.storageGroups[0].stopNodes(tc, 1)
assertSeries(opts.c.vmselect, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnFullResponse)
// Stop one more vmstoarge in remaining globarRF-1 groups.
//
@@ -1089,19 +1096,56 @@ func testGroupPartialResponse(tc *apptest.TestCase, opts *testGroupReplicationOp
for g := 1; g < opts.globalRF-1; g++ {
opts.c.storageGroups[g].stopNodes(tc, 1)
}
assertSeries(opts.c.vmselect, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnFullResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnFullResponse)
// Stop one more vmstoarge in one more group.
//
// vmselectGroupGlobalRF must now return partial dataset.
opts.c.storageGroups[opts.globalRF].stopNodes(tc, 1)
assertSeries(opts.c.vmselect, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, mustReturnPartialResponse)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnPartialResponse)
// Stop all the remaining vmstorage nodes except a single node.
//
// At this point vmselects still must be able to return partial response
// because at least one vmstorage node has successfully returned results.
n := len(opts.c.storageGroups[0].vmstorages)
opts.c.storageGroups[0].stopNodes(tc, n-1)
for g := 1; g < len(opts.c.storageGroups); g++ {
n := len(opts.c.storageGroups[g].vmstorages)
opts.c.storageGroups[g].stopNodes(tc, n)
}
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnPartialResponse)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnPartialResponse)
mustReturnUnavailableError := &apptest.PrometheusAPIV1SeriesResponse{
Status: "error",
ErrorType: "503",
}
// vmselects must return an error for the same request when partial
// responses are denied explicitly.
assertSeries(opts.c.vmselect, denyPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGroupRF, denyPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGlobalRF, denyPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGroupGlobalRF, denyPartialResponse, mustReturnUnavailableError)
// Stop the last remaining vmstorage node.
//
// vmselects must return an error when there are no successful vmstorage
// responses.
opts.c.storageGroups[0].stopNodes(tc, 1)
assertSeries(opts.c.vmselect, allowPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGroupRF, allowPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGlobalRF, allowPartialResponse, mustReturnUnavailableError)
assertSeries(opts.c.vmselectGroupGlobalRF, allowPartialResponse, mustReturnUnavailableError)
}
// TestClusterReplication_PartialResponseMultitenant checks how vmselect handles some

View File

@@ -0,0 +1,51 @@
{
"ulid": "01JFJBS3YP1SHZ3PJQ6HK76EC3",
"minTime": 1734709200000,
"maxTime": 1734709320000,
"stats": {
"numSamples": 400,
"numSeries": 100,
"numChunks": 100
},
"compaction": {
"level": 1,
"sources": [
"01JFJBS3YP1SHZ3PJQ6HK76EC3"
],
"parents": [
{
"ulid": "00000000000000000000000000",
"minTime": 0,
"maxTime": 0
}
],
"hints": [
"from-out-of-order"
]
},
"version": 1,
"out_of_order": false,
"thanos": {
"labels": {},
"downsample": {
"resolution": 0
},
"source": "receive",
"segment_files": [
"000001"
],
"files": [
{
"rel_path": "chunks/000001",
"size_bytes": 4808
},
{
"rel_path": "index",
"size_bytes": 55021
},
{
"rel_path": "meta.json"
}
]
}
}

Binary file not shown.

File diff suppressed because one or more lines are too long

View File

@@ -0,0 +1,139 @@
package tests
import (
"encoding/json"
"fmt"
"io"
"os"
"testing"
"github.com/google/go-cmp/cmp"
"github.com/google/go-cmp/cmp/cmpopts"
"github.com/VictoriaMetrics/VictoriaMetrics/apptest"
"github.com/VictoriaMetrics/VictoriaMetrics/lib/fs"
)
const (
testMimirPath = "testdata/mimir-tsdb"
expectedMimirResponseFile = "./testdata/mimir-tsdb/expected_response.json"
)
func TestSingleVmctlMimirProtocol(t *testing.T) {
fs.MustRemoveDir(t.Name())
tc := apptest.NewTestCase(t)
defer tc.Stop()
vmsingleDst := tc.MustStartDefaultVmsingle()
vmAddr := fmt.Sprintf("http://%s/", vmsingleDst.HTTPAddr())
dir, err := os.Getwd()
if err != nil {
t.Fatalf("cannot get current working directory: %s", err)
}
path := fmt.Sprintf("fs://%s/%s", dir, testMimirPath)
vmctlFlags := []string{
`mimir`,
`--mimir-tenant-id=anonymous`,
`--mimir-filter-time-start=2024-12-01T00:00:00Z`,
`--mimir-filter-time-end=2024-12-31T23:59:59Z`,
`--mimir-custom-s3-endpoint=http://localhost:9000`,
`--mimir-path=` + path,
`--vm-addr=` + vmAddr,
`--disable-progress-bar=true`,
`--vm-concurrency=6`,
`--mimir-concurrency=6`,
}
testMimirProtocol(tc, vmsingleDst, vmctlFlags)
}
func TestClusterVmctlMimirProtocol(t *testing.T) {
fs.MustRemoveDir(t.Name())
tc := apptest.NewTestCase(t)
defer tc.Stop()
cluster := tc.MustStartDefaultCluster()
vmAddr := fmt.Sprintf("http://%s/", cluster.Vminsert.HTTPAddr())
dir, err := os.Getwd()
if err != nil {
t.Fatalf("cannot get current working directory: %s", err)
}
path := fmt.Sprintf("fs://%s/%s", dir, testMimirPath)
vmctlFlags := []string{
`mimir`,
`--mimir-tenant-id=anonymous`,
`--mimir-filter-time-start=2024-12-01T00:00:00Z`,
`--mimir-filter-time-end=2024-12-31T23:59:59Z`,
`--mimir-custom-s3-endpoint=http://localhost:9000`,
`--mimir-path=` + path,
`--vm-addr=` + vmAddr,
`--disable-progress-bar=true`,
`--vm-concurrency=6`,
`--mimir-concurrency=6`,
}
testMimirProtocol(tc, cluster, vmctlFlags)
}
func testMimirProtocol(tc *apptest.TestCase, sut apptest.PrometheusWriteQuerier, vmctlFlags []string) {
t := tc.T()
t.Helper()
cmpOpt := cmpopts.IgnoreFields(apptest.PrometheusAPIV1QueryResponse{}, "Status", "Data.ResultType")
// test for empty data request
got := sut.PrometheusAPIV1Query(t, `{__name__=~".*"}`, apptest.QueryOpts{
Step: "5m",
Time: "2025-06-02T17:14:00Z",
})
want := apptest.NewPrometheusAPIV1QueryResponse(t, `{"data":{"result":[]}}`)
if diff := cmp.Diff(want, got, cmpOpt); diff != "" {
t.Errorf("unexpected response (-want, +got):\n%s", diff)
}
tc.MustStartVmctl("vmctl", vmctlFlags)
sut.ForceFlush(t)
// open the expected series response file
file, err := os.Open(expectedMimirResponseFile)
if err != nil {
t.Fatalf("cannot open expected series response file: %s", err)
}
defer file.Close()
bytes, err := io.ReadAll(file)
if err != nil {
t.Fatalf("cannot read expected series response file: %s", err)
}
var wantResponse apptest.PrometheusAPIV1QueryResponse
if err := json.Unmarshal(bytes, &wantResponse); err != nil {
t.Fatalf("cannot unmarshal expected series response file: %s", err)
}
wantResponse.Sort()
tc.Assert(&apptest.AssertOptions{
// For cluster version, we need to wait longer for the metrics to be stored
Retries: 300,
Msg: `unexpected metrics stored on vmsingle via the prometheus protocol`,
Got: func() any {
expected := sut.PrometheusAPIV1Export(t, `{__name__=~".*"}`, apptest.QueryOpts{
Start: "2024-12-01T15:31:10Z",
End: "2024-12-31T15:32:20Z",
})
expected.Sort()
return expected.Data.Result
},
Want: wantResponse.Data.Result,
CmpOpts: []cmp.Option{
cmpopts.IgnoreFields(apptest.PrometheusAPIV1QueryResponse{}, "Status", "Data.ResultType"),
},
})
}

View File

@@ -130,7 +130,7 @@
"calcs": [
"lastNotNull"
],
"fields": "/^short_version$/",
"fields": "/^version$/",
"values": false
},
"showPercentChange": false,
@@ -146,11 +146,10 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "vm_app_version{job=~\"$job\",instance=~\"$instance\"}",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"interval": "",
"legendFormat": "{{short_version}}",
"range": false,
"refId": "A"
}

View File

@@ -791,7 +791,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -789,7 +789,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -131,7 +131,7 @@
"calcs": [
"lastNotNull"
],
"fields": "/^short_version$/",
"fields": "/^version$/",
"values": false
},
"showPercentChange": false,
@@ -147,11 +147,10 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "vm_app_version{job=~\"$job\",instance=~\"$instance\"}",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"interval": "",
"legendFormat": "{{short_version}}",
"range": false,
"refId": "A"
}

View File

@@ -792,7 +792,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -790,7 +790,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -785,7 +785,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -566,7 +566,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,
@@ -2804,10 +2804,10 @@
"overrides": []
},
"gridPos": {
"h": 8,
"h": 7,
"w": 12,
"x": 0,
"y": 352
"y": 11
},
"id": 63,
"options": {
@@ -2843,7 +2843,113 @@
],
"title": "Restarts ($job)",
"type": "timeseries"
}
},
{
"datasource": {
"type": "victoriametrics-metrics-datasource",
"uid": "$ds"
},
"description": "Group iteration reset can be caused by irregular delays during evaluation or by the system wall clock being moved backward.\nIf it is caused by host clock changes, vmalert could generate duplicate results for the group rules, since some evaluations could be repeated.\nCheck the host clock time synchronization configuration if this happens frequently.\n",
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "bars",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "linear",
"lineWidth": 1,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 80
}
]
},
"unit": "short"
},
"overrides": []
},
"gridPos": {
"h": 7,
"w": 12,
"x": 12,
"y": 11
},
"id": 70,
"options": {
"legend": {
"calcs": [
"mean",
"lastNotNull",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "none"
}
},
"pluginVersion": "12.2.0",
"targets": [
{
"datasource": {
"type": "victoriametrics-metrics-datasource",
"uid": "$ds"
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(increase(vmalert_iteration_reset_total{job=~\"$job\", instance=~\"$instance\"}[$__rate_interval])) by(job, group, file) > 0",
"interval": "1m",
"legendFormat": "({{job}}) {{group}}({{file}})",
"range": true,
"refId": "A"
}
],
"title": "Group Iteration Reset ($instance)",
"type": "timeseries"
}
],
"title": "Troubleshooting",
"type": "row"

View File

@@ -492,14 +492,14 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by (job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,
"refId": "A"
}
],
"title": "Version",
"title": "",
"type": "table"
},
{

View File

@@ -784,7 +784,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,

View File

@@ -565,7 +565,7 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by(job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,
@@ -2803,10 +2803,10 @@
"overrides": []
},
"gridPos": {
"h": 8,
"h": 7,
"w": 12,
"x": 0,
"y": 352
"y": 11
},
"id": 63,
"options": {
@@ -2842,7 +2842,113 @@
],
"title": "Restarts ($job)",
"type": "timeseries"
}
},
{
"datasource": {
"type": "prometheus",
"uid": "$ds"
},
"description": "Group iteration reset can be caused by irregular delays during evaluation or by the system wall clock being moved backward.\nIf it is caused by host clock changes, vmalert could generate duplicate results for the group rules, since some evaluations could be repeated.\nCheck the host clock time synchronization configuration if this happens frequently.\n",
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "bars",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "linear",
"lineWidth": 1,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 80
}
]
},
"unit": "short"
},
"overrides": []
},
"gridPos": {
"h": 7,
"w": 12,
"x": 12,
"y": 11
},
"id": 70,
"options": {
"legend": {
"calcs": [
"mean",
"lastNotNull",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "none"
}
},
"pluginVersion": "12.2.0",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "$ds"
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(increase(vmalert_iteration_reset_total{job=~\"$job\", instance=~\"$instance\"}[$__rate_interval])) by(job, group, file) > 0",
"interval": "1m",
"legendFormat": "({{job}}) {{group}}({{file}})",
"range": true,
"refId": "A"
}
],
"title": "Group Iteration Reset ($instance)",
"type": "timeseries"
}
],
"title": "Troubleshooting",
"type": "row"

View File

@@ -491,14 +491,14 @@
},
"editorMode": "code",
"exemplar": false,
"expr": "sum(vm_app_version{job=~\"$job\", instance=~\"$instance\"}) by (job, short_version)",
"expr": "sum by(job, version) (label_replace(vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version!=\"\"}, \"version\", \"$1\", \"short_version\", \"(.*)\") or vm_app_version{job=~\"$job\", instance=~\"$instance\", short_version=\"\"})",
"format": "table",
"instant": true,
"range": false,
"refId": "A"
}
],
"title": "Version",
"title": "",
"type": "table"
},
{

View File

@@ -7,7 +7,7 @@ ROOT_IMAGE ?= alpine:3.23.4
ROOT_IMAGE_SCRATCH ?= scratch
CERTS_IMAGE := alpine:3.23.4
GO_BUILDER_IMAGE := golang:1.26.3
GO_BUILDER_IMAGE := golang:1.26.4
BUILDER_IMAGE := local/builder:2.0.0-$(shell echo $(GO_BUILDER_IMAGE) | tr :/ __)-1
BASE_IMAGE := local/base:1.1.4-$(shell echo $(ROOT_IMAGE) | tr :/ __)-$(shell echo $(CERTS_IMAGE) | tr :/ __)

View File

@@ -64,6 +64,18 @@ groups:
group \"{{ $labels.group }}\". See https://docs.victoriametrics.com/victoriametrics/vmalert/#groups.
If rule expressions are taking longer than expected, please see https://docs.victoriametrics.com/victoriametrics/troubleshooting/#slow-queries."
- alert: GroupIterationReset
expr: increase(vmalert_iteration_reset_total[5m]) > 0
for: 5m
labels:
severity: warning
annotations:
summary: "Evaluation iteration for group {{ $labels.group }} in file {{ $labels.file }} is reset"
description: "Evaluation iteration for group \"{{ $labels.group }}\" in file \"{{ $labels.file }}\" is reset on vmalert instance {{ $labels.instance }}.
This can be caused by irregular delays during evaluation or by the system wall clock being moved backward. If it is caused by host clock changes, vmalert could
generate duplicate results for the group rules since some evaluations could be repeated. Check host clock time synchronization configurations if this happens frequently."
- alert: RemoteWriteErrors
expr: increase(vmalert_remotewrite_errors_total[5m]) > 0
for: 15m
@@ -108,4 +120,3 @@ groups:
summary: "vmalert instance {{ $labels.instance }} is failing to send notifications to Alertmanager"
description: "vmalert instance {{ $labels.instance }} is failing to send alert notifications to \"{{ $labels.addr }}\".
Check vmalert's logs for detailed error message."

View File

@@ -130,7 +130,7 @@ Released: 2025-11-05
## v1.27.0
Released: 2025-10-31
- FEATURE: Added runtime state compatibility guard for [stateful](https://docs.victoriametrics.com/anomaly-detection/components/settings/#restore-state) deployments. The service now persists normalized versions, evaluates an [upgrade/downgrade compatibility matrix](https://docs.victoriametrics.com/anomaly-detection/migration/#compatibility-matrix), and selectively drops or reuses DB records and on-disk artifacts to keep migrations safe and automatic. Please refer to the [migration page](https://docs.victoriametrics.com/anomaly-detection/migration/) for more details.
- FEATURE: Added runtime state compatibility guard for [stateful](https://docs.victoriametrics.com/anomaly-detection/components/settings/#state-restoration) deployments. The service now persists normalized versions, evaluates an [upgrade/downgrade compatibility matrix](https://docs.victoriametrics.com/anomaly-detection/migration/#compatibility-matrix), and selectively drops or reuses DB records and on-disk artifacts to keep migrations safe and automatic. Please refer to the [migration page](https://docs.victoriametrics.com/anomaly-detection/migration/) for more details.
- IMPROVEMENT: Parallelization now honours container cgroup CPU/RAM limits, so `settings.n_workers` in the [settings section](https://docs.victoriametrics.com/anomaly-detection/components/settings/#parallelization), internal routines and the `vmanomaly_available_memory_bytes`/`vmanomaly_cpu_cores_available` [startup metrics](https://docs.victoriametrics.com/anomaly-detection/components/monitoring/#startup-metrics) report or use container resources instead of host totals, keeping the [self-monitoring dashboard](https://docs.victoriametrics.com/anomaly-detection/self-monitoring/#grafana-dashboard) accurate.
@@ -190,7 +190,7 @@ Released: 2025-08-19
## v1.25.2
Released: 2025-07-30
- BUGFIX: Resolved inconsistent state between in-memory models and state database (if [stateful mode](https://docs.victoriametrics.com/anomaly-detection/components/settings/#stateful-mode) is enabled). This bug caused `Model instance not found` warnings during inference calls and prevented proper cleanup of stale models from disk. The fix also prevents state updates when operations are terminated mid-execution of scheduled fit/infer jobs.
- BUGFIX: Resolved inconsistent state between in-memory models and state database (if [stateful mode](https://docs.victoriametrics.com/anomaly-detection/components/settings/#state-restoration) is enabled). This bug caused `Model instance not found` warnings during inference calls and prevented proper cleanup of stale models from disk. The fix also prevents state updates when operations are terminated mid-execution of scheduled fit/infer jobs.
- BUGFIX: Added explicit handling for inference calls on models that were deleted from disk by the time of their usage, but still referenced in the state database, preventing `'NoneType' object has no attribute 'infer'` rows in logs. Now a warning is logged and the inference call is skipped, which is expected behavior for deleted models.
@@ -210,7 +210,7 @@ Released: 2025-07-24
- BUGFIX: Prevented `OneOffScheduler` and `BacktestingScheduler` [schedulers](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/) from receiving no data (when [state restoration](https://docs.victoriametrics.com/anomaly-detection/components/settings/#state-restoration) is enabled). Now a warning is logged and such scheduler types are implicitly used without state restoration, which is expected behavior for these one-time-job schedulers.
- BUGFIX: Now the paths to artifact database (if [stateful mode](https://docs.victoriametrics.com/anomaly-detection/components/settings/#stateful-mode) is enabled) are properly resolved to absolute, preventing errors at initialization time (like `sqlalchemy.exc.OperationalError: (sqlite3.OperationalError) unable to open database file`) or warnings (like `SAWarning: fully NULL primary key identity cannot load any object.`).
- BUGFIX: Now the paths to artifact database (if [stateful mode](https://docs.victoriametrics.com/anomaly-detection/components/settings/#state-restoration) is enabled) are properly resolved to absolute, preventing errors at initialization time (like `sqlalchemy.exc.OperationalError: (sqlite3.OperationalError) unable to open database file`) or warnings (like `SAWarning: fully NULL primary key identity cannot load any object.`).
## v1.25.0
Released: 2025-07-17
@@ -559,7 +559,7 @@ Released: 2024-08-10
- **Lowest anomaly scores** (=0) when the *model's predictions (`yhat`) fall outside the expected range*, signaling uncertain predictions.
- For more details, please refer to the [documentation](https://docs.victoriametrics.com/anomaly-detection/components/reader/#per-query-parameters).
- IMPROVEMENT: Added `latency_offset` argument to the [VmReader](https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader) to override the default `-search.latencyOffset` [flag of VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags) (30s). The default value is set to 1ms, which should help in cases where `sampling_frequency` is low (10-60s) and `sampling_frequency` equals `infer_every` in the [PeriodicScheduler](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). This prevents users from receiving `service - WARNING - [Scheduler [scheduler_alias]] No data available for inference.` warnings in logs and allows for consecutive `infer` calls without gaps. To restore the backward compatible behavior, set it equal to your `-search.latencyOffset` value in [VmReader](https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader) config section.
- IMPROVEMENT: Added `latency_offset` argument to the [VmReader](https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader) to override the default `-search.latencyOffset` [flag of VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags) (30s). The default value is set to 1ms, which should help in cases where `sampling_period` is low (10-60s) and `sampling_period` equals `infer_every` in the [PeriodicScheduler](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). This prevents users from receiving `service - WARNING - [Scheduler [scheduler_alias]] No data available for inference.` warnings in logs and allows for consecutive `infer` calls without gaps. To restore the backward compatible behavior, set it equal to your `-search.latencyOffset` value in [VmReader](https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader) config section.
- BUGFIX: Ensure the `use_transform` argument of the [`OnlineQuantileModel`](https://docs.victoriametrics.com/anomaly-detection/components/models/#online-seasonal-quantile) functions as intended.
- BUGFIX: Add a docstring for `query_from_last_seen_timestamp` arg of [VmReader](https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader).

View File

@@ -281,7 +281,7 @@ reader:
datasource_url: 'some_url_to_read_data_from'
queries:
query_alias1: 'some_metricsql_query'
sampling_frequency: '1m' # change to whatever you need in data granularity
sampling_period: '1m' # change to whatever you need in data granularity
# other params if needed
# https://docs.victoriametrics.com/anomaly-detection/components/reader/#vm-reader
@@ -294,7 +294,7 @@ writer:
# https://docs.victoriametrics.com/anomaly-detection/components/monitoring/
```
Configuration above will produce N intervals of full length (`fit_window`=14d + `fit_every`=1h) until `to_iso` timestamp is reached to run N consecutive `fit` calls to train models; Then these models will be used to produce `M = [fit_every / sampling_frequency]` infer datapoints for `fit_every` range at the end of each such interval, imitating M consecutive calls of `infer_every` in `PeriodicScheduler` [config](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). These datapoints then will be written back to VictoriaMetrics TSDB, defined in `writer` [section](https://docs.victoriametrics.com/anomaly-detection/components/writer/#vm-writer) for further visualization (i.e. in VMUI or Grafana)
Configuration above will produce N intervals of full length (`fit_window`=14d + `fit_every`=1h) until `to_iso` timestamp is reached to run N consecutive `fit` calls to train models; Then these models will be used to produce `M = [fit_every / sampling_period]` infer datapoints for `fit_every` range at the end of each such interval, imitating M consecutive calls of `infer_every` in `PeriodicScheduler` [config](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). These datapoints then will be written back to VictoriaMetrics TSDB, defined in `writer` [section](https://docs.victoriametrics.com/anomaly-detection/components/writer/#vm-writer) for further visualization (i.e. in VMUI or Grafana)
## Forecasting
@@ -499,7 +499,7 @@ schedulers:
models:
zscore_example:
class: 'zscore_online'
min_n_samples_seen: 120 # i.e. minimal relevant seasonality or (initial) fit_window / sampling_frequency
min_n_samples_seen: 120 # i.e. minimal relevant seasonality or (initial) fit_window / sampling_period
decay: 0.999 # decay factor to control how fast the model adapts to new data, the lower, the faster it adapts
schedulers: ['periodic']
# other model params ...

View File

@@ -41,7 +41,7 @@ settings:
restore_state: True # restore state from previous run, if available
retention: # how long to keep stale models on disk/in memory
ttl: "1d" # time-to-live duration, if the model was not used for inference within this duration, it will be considered stale
check_every: "1h" # how often to check for stale models and remove them
check_interval: "1h" # how often to check for stale models and remove them
# how and when to run the models is defined by schedulers
# https://docs.victoriametrics.com/anomaly-detection/components/scheduler/
@@ -143,11 +143,11 @@ server:
> This feature is better used in conjunction with [stateful service](https://docs.victoriametrics.com/anomaly-detection/components/settings/#state-restoration) to preserve the state of the models and schedulers between restarts and reuse what can be reused, thus avoiding unnecessary re-training of models, re-initialization of schedulers and re-reading of data.
{{% available_from "v1.25.0" anomaly %}} Service supports hot reload of configuration files, which allows for automatic reloading of configurations on config files change filesystem events without the need of explicit service restart. This can be enabled via the `--watch` [CLI argument](https://docs.victoriametrics.com/anomaly-detection/quickstart/#command-line-arguments). `vmanomaly_hot_reload_enabled` flag in [self-monitoring metrics](https://docs.victoriametrics.com/anomaly-detection/components/monitoring/#startup-metrics) will be set to 1 (if enabled) or 0 (if disabled).
{{% available_from "v1.25.0" anomaly %}} Service supports hot reload of configuration files, which allows for automatic reloading of configurations on config files change filesystem events without the need of explicit service restart. This can be enabled via the `--watch` [CLI argument](https://docs.victoriametrics.com/anomaly-detection/quickstart/#command-line-arguments). `vmanomaly_config_reload_enabled` flag in [self-monitoring metrics](https://docs.victoriametrics.com/anomaly-detection/components/monitoring/#startup-metrics) will be set to 1 (if enabled) or 0 (if disabled).
### How it works
It works by watching for file system events, such as modifications, creations, or deletions of `.yml|.yaml` files in the specified directories. When a change is detected, the service will attempt to reload the configuration files, rebuild the [global config](https://docs.victoriametrics.com/anomaly-detection/scaling-vmanomaly/#global-config) and reinitialize the components. If the reload is successful, the `vmanomaly_hot_reload_events_total` metric will be incremented for `status="success"` label, otherwise it will be incremented with `status="failure"` label and a respective error message on config validation failure(s) will be logged.
It works by watching for file system events, such as modifications, creations, or deletions of `.yml|.yaml` files in the specified directories. When a change is detected, the service will attempt to reload the configuration files, rebuild the [global config](https://docs.victoriametrics.com/anomaly-detection/scaling-vmanomaly/#global-config) and reinitialize the components. If the reload is successful, the `vmanomaly_config_reloads_total` metric will be incremented for `status="success"` label, otherwise it will be incremented with `status="failure"` label and a respective error message on config validation failure(s) will be logged.
> If the reload fails, the service will log an error message indicating the reason for the failure, and the **previous configuration will remain active until a successful reload occurs** to preserve the service's stability. This means that if there are errors in the new configuration, the service will continue to operate with the last valid configuration until the issues are resolved.
@@ -219,7 +219,7 @@ reader:
# ... (rest of the config remains unchanged)
```
After saving the changes, hot reload will automatically detect the changes in `config.yaml` and attempt to reload the configuration. As the changes are valid, the service will log a success message and increment the `vmanomaly_hot_reload_events_total` metric with `status="success"` label:
After saving the changes, hot reload will automatically detect the changes in `config.yaml` and attempt to reload the configuration. As the changes are valid, the service will log a success message and increment the `vmanomaly_config_reloads_total` metric with `status="success"` label:
- All the model instances of class `zscore_online`, that were trained on `host_network_receive_errors` can be reused as they are still valid and "fresh" for making inference on new datapoints until the next `fit_every` happens.
- All the model instances of class `zscore_online`, that were trained on `cpu_seconds_total` will be re-trained with the new query expression and frequency, as old model instances are not valid anymore.

View File

@@ -369,7 +369,7 @@ If True, then query will be performed from the last seen timestamp for a given s
`1ms`
</td>
<td>
It allows overriding the default `-search.latencyOffset`{{% available_from "v1.15.1" anomaly %}} [flag of VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags) (30s). The default value is set to 1ms, which should help in cases where `sampling_frequency` is low (10-60s) and `sampling_frequency` equals `infer_every` in the [PeriodicScheduler](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). This prevents users from receiving `service - WARNING - [Scheduler [scheduler_alias]] No data available for inference.` warnings in logs and allows for consecutive `infer` calls without gaps. To restore the old behavior, set it equal to your `-search.latencyOffset` [flag value](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags).
It allows overriding the default `-search.latencyOffset`{{% available_from "v1.15.1" anomaly %}} [flag of VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags) (30s). The default value is set to 1ms, which should help in cases where `sampling_period` is low (10-60s) and `sampling_period` equals `infer_every` in the [PeriodicScheduler](https://docs.victoriametrics.com/anomaly-detection/components/scheduler/#periodic-scheduler). This prevents users from receiving `service - WARNING - [Scheduler [scheduler_alias]] No data available for inference.` warnings in logs and allows for consecutive `infer` calls without gaps. To restore the old behavior, set it equal to your `-search.latencyOffset` [flag value](https://docs.victoriametrics.com/victoriametrics/#list-of-command-line-flags).
</td>
</tr>
<tr>

View File

@@ -6,342 +6,45 @@ build:
sitemap:
disable: true
---
**Objective**
[VictoriaMetrics Enterprise](https://docs.victoriametrics.com/victoriametrics/enterprise/) supports specifying multiple retentions for distinct sets of time series and tenants. If you are an Enterprise user, [configure multiple retentions directly through retention filters](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#retention-filters) instead of following this guide.
Setup Victoria Metrics Cluster with support of multiple retention periods within one installation.
This guide explains how to set up multiple retentions using an [open-source VictoriaMetrics Cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/).
**Enterprise Solution**
## Overview
[VictoriaMetrics Enterprise](https://docs.victoriametrics.com/victoriametrics/enterprise/) supports specifying multiple retentions
for distinct sets of time series and [tenants](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#multitenancy)
via [retention filters](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#retention-filters).
VictoriaMetrics retains metrics by default for 1 month. You can change data retention with the [`-retentionPeriod` command-line flag](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/#retention), but this value applies to all time series stored on a given `vmstorage` node and cannot be customized per tenant or per metric in the open source version.
**Open Source Solution**
## Multi-Retention Architecture
Community version of VictoriaMetrics supports only one retention period per `vmstorage` node via [-retentionPeriod](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/#retention) command-line flag.
To support multiple retentions with the open source version of VictoriaMetrics cluster, you can split the cluster into several logical groups of `vmstorage` nodes, where each group is configured with a different `-retentionPeriod` and receives only the data that must follow that retention.
A multi-retention setup can be implemented by dividing a [victoriametrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/) into logical groups with different retentions.
Each storage group is connected to a separate `vminsert`, while a shared `vmselect` layer queries across all storage groups so that dashboards and alerts continue to see a single logical VictoriaMetrics backend.
Example:
Setup should handle 3 different retention groups 3months, 1year and 3 years.
Solution contains 3 groups of vmstorages + vminserts and one group of vmselects. Routing is done by [vmagent](https://docs.victoriametrics.com/victoriametrics/vmagent/)
by [splitting data streams](https://docs.victoriametrics.com/victoriametrics/vmagent/#splitting-data-streams-among-multiple-systems).
The [-retentionPeriod](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/#retention) sets how long to keep the metrics.
The diagram below shows a proposed solution
![Setup](setup.webp)
In the example used throughout this guide, the cluster is divided into three groups:
**Implementation Details**
- Group A: 3-month retention.
- Group B: 1-year retention.
- Group C: 3-year retention.
1. Groups of vminserts A know about only vmstorages A and this is explicitly specified via `-storageNode` [configuration](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#cluster-setup).
1. Groups of vminserts B know about only vmstorages B and this is explicitly specified via `-storageNode` [configuration](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#cluster-setup).
1. Groups of vminserts C know about only vmstorages C and this is explicitly specified via `-storageNode` [configuration](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#cluster-setup).
1. vmselect reads data from all vmstorage nodes via `-storageNode` [configuration](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#cluster-setup)
with [deduplication](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#deduplication) setting equal to vmagent's scrape interval or minimum interval between collected samples.
1. vmagent routes incoming metrics to the given set of `vminsert` nodes using relabeling rules specified at `-remoteWrite.urlRelabelConfig` [configuration](https://docs.victoriametrics.com/victoriametrics/relabeling/).
Metrics are routed to the appropriate `vminsert` group by [splitting data streams](https://docs.victoriametrics.com/victoriametrics/vmagent/#splitting-data-streams-among-multiple-systems) in `vmagent`. An optional [vmauth](https://docs.victoriametrics.com/victoriametrics/vmauth/) rule can be added on top to enforce per-tenant routing or API access policies.
**Multi-Tenant Setup**
## Implementing Multi-Retention on Kubernetes
In this section, we'll install and configure the components for a multi-retention deployment of the VictoriaMetrics cluster. See [Kubernetes monitoring with VictoriaMetrics Cluster](https://docs.victoriametrics.com/guides/k8s-monitoring-via-vm-cluster/) for details on running VictoriaMetrics in Kubernetes.
Run the following command to add the VictoriaMetrics Helm repository:
```shell
helm repo add vm https://victoriametrics.github.io/helm-charts/
helm repo update
```
### Step 1: Deploying storage groups
We'll create three retention groups. Each has a different retention period and disk size. Read [Understand Your Setup Size](https://docs.victoriametrics.com/guides/understand-your-setup-size/) to estimate how much space you will need for each group. The following table is shown as an example.
| Group | Retention Period | Disk Size |
|-------------|------------------|-----------|
| `vmcluster-a` | 3 months (`3M`) | 80GB |
| `vmcluster-b` | 1 year (`1Y`) | 300 GB |
| `vmcluster-c` | 3 years (`3Y`) | 900 GB |
Create a Helm values file for Group A.
```shell
cat <<EOF > vmcluster-a.yaml
vmstorage:
enabled: true
persistence:
size: 80Gi
extraArgs:
retentionPeriod: 3M
storageDataPath: /vmstorage-data
dedup.minScrapeInterval: 30s
podLabels:
retention-group: a
retention-period: 3M
vminsert:
enabled: true
podLabels:
retention-group: a
vmselect:
enabled: false
EOF
```
The values file above creates `vminsert` and `vmstorage` services while turning off `vmselect`, which we'll deploy separately. It also defines a 30-second [deduplication](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#deduplication) window to handle possible duplicate metrics. The deduplication window must match the `vmagent` service scrape window (which we'll define later in the guide).
Create the values files for Group B and Group C:
```shell
cat <<EOF > vmcluster-b.yaml
vmstorage:
enabled: true
persistence:
size: 300Gi
extraArgs:
retentionPeriod: 1Y
storageDataPath: /vmstorage-data
dedup.minScrapeInterval: 30s
podLabels:
retention-group: b
retention-period: 1Y
vminsert:
enabled: true
podLabels:
retention-group: b
vmselect:
enabled: false
EOF
cat <<EOF > vmcluster-c.yaml
vmstorage:
enabled: true
persistence:
size: 900Gi
extraArgs:
retentionPeriod: 3Y
storageDataPath: /vmstorage-data
dedup.minScrapeInterval: 30s
podLabels:
retention-group: c
retention-period: 3Y
vminsert:
enabled: true
podLabels:
retention-group: c
vmselect:
enabled: false
EOF
```
Deploy the three storage groups with:
```shell
helm upgrade --install vmcluster-a vm/victoria-metrics-cluster -f vmcluster-a.yaml
helm upgrade --install vmcluster-b vm/victoria-metrics-cluster -f vmcluster-b.yaml
helm upgrade --install vmcluster-c vm/victoria-metrics-cluster -f vmcluster-c.yaml
# Wait for all storage pods to be ready
kubectl rollout status statefulset -l app.kubernetes.io/instance=vmcluster-a
kubectl rollout status statefulset -l app.kubernetes.io/instance=vmcluster-b
kubectl rollout status statefulset -l app.kubernetes.io/instance=vmcluster-c
```
### Step 2: Deploying vmselect
Next, we'll deploy a `vmselect` service to route queries to the storage groups.
Create a Helm values file with:
```shell
cat <<EOF >vmselect.yaml
vmstorage:
enabled: false
vminsert:
enabled: false
vmselect:
enabled: true
replicaCount: 2
suppressStorageFQDNsRender: true
extraArgs:
# Each list item is a single -storageNode flag with comma-separated hosts
# in the same group. The FQDN format is:
# <pod>.<svc>.default.svc
# where pod = <release>-victoria-metrics-cluster-vmstorage-<N>
# and svc = <release>-victoria-metrics-cluster-vmstorage
storageNode:
- "a/vmcluster-a-victoria-metrics-cluster-vmstorage-0.vmcluster-a-victoria-metrics-cluster-vmstorage.default.svc:8401,a/vmcluster-a-victoria-metrics-cluster-vmstorage-1.vmcluster-a-victoria-metrics-cluster-vmstorage.default.svc:8401"
- "b/vmcluster-b-victoria-metrics-cluster-vmstorage-0.vmcluster-b-victoria-metrics-cluster-vmstorage.default.svc:8401,b/vmcluster-b-victoria-metrics-cluster-vmstorage-1.vmcluster-b-victoria-metrics-cluster-vmstorage.default.svc:8401"
- "c/vmcluster-c-victoria-metrics-cluster-vmstorage-0.vmcluster-c-victoria-metrics-cluster-vmstorage.default.svc:8401,c/vmcluster-c-victoria-metrics-cluster-vmstorage-1.vmcluster-c-victoria-metrics-cluster-vmstorage.default.svc:8401"
dedup.minScrapeInterval: 30s
EOF
```
Let's break down the file above:
- Deploys `vmselect` as a separate Helm release
- Disables `vminsert` and `vmstorage` as these services were already deployed in Step 1.
- `supressStorageFQDNsRender: true` turns off automatic FQDN generation for storage nodes. By default, the Helm chart auto-generates `-storageNodes` flags, but since `vmstorage` has been disabled, we need to supply them manually in `extraArgs`.
- In `extraArgs.storageNode:` we define the list of `vmstorage` services to reach for queries. The `storageNode` flags tell vmselect to query all 6 storage pods, which are organized into three groups: `a`, `b`, and `c`.
Deploy the `vmselect` release with:
```shell
helm upgrade --install vmselect vm/victoria-metrics-cluster -f vmselect.yaml
```
### Step 3: Deploying vmagent
We'll use `vmagent` to route incoming metrics to the correct retention group. For example, we can use a `retention` label for mapping metrics to storage groups in the following way:
| `retention` label | Storage Group |
| ----------------| --------------|
| `"3mo"` | `vmcluster-a` |
| `"1yr"` | `vmcluster-b` |
| `"3yr"` | `vmcluster-c` |
Create the values file for vmagent:
```shell
cat <<EOF >vmagent.yaml
service:
enabled: true
remoteWrite:
# Group A: receives metrics with retention="3mo"
- url: http://vmcluster-a-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write
urlRelabelConfig:
- action: keep
source_labels: [retention]
regex: "3mo"
# Group B: receives metrics with retention="1yr"
- url: http://vmcluster-b-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write
urlRelabelConfig:
- action: keep
source_labels: [retention]
regex: "1yr"
# Group C: receives metrics with retention="3yr"
- url: http://vmcluster-c-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write
urlRelabelConfig:
- action: keep
source_labels: [retention]
regex: "3yr"
EOF
```
> Two important notes on scraping:
> - Metrics without a matching `retention` label are silently dropped by the `keep` rules. You must ensure that every metric receives a label, or use a different routing configuration.
> - The [scrape interval](https://docs.victoriametrics.com/victoriametrics/sd_configs/#scrape_configs) should match the `dedup.minScrapeInterval` defined in the vmstorage nodes.
`
Now deploy the vmagent release:
```shell
helm upgrade --install vmagent vm/victoria-metrics-agent -f vmagent.yaml
```
Wait for vmagent to become ready:
```shell
kubectl rollout status deploy/vmagent-victoria-metrics-agent
```
### Step 4: Verification
We can send test data to verify that data is flowing to the correct storage group.
First, port-forward vmagent and vmselect:
```shell
VMAGENT_SVC=$(kubectl get svc -l app.kubernetes.io/instance=vmagent -o jsonpath='{.items[0].metadata.name}')
kubectl port-forward "svc/$VMAGENT_SVC" 8429 &
VMSELECT_SVC=$(kubectl get svc -l app.kubernetes.io/instance=vmselect -o jsonpath='{.items[0].metadata.name}')
kubectl port-forward "svc/$VMSELECT_SVC" 8481 &
```
Send test metrics directly to vmagent's HTTP endpoint to exercise all three retention labels:
```shell
POD=$(kubectl get pod -l app.kubernetes.io/instance=vmagent -o jsonpath='{.items[0].metadata.name}')
for retention in 3mo 1yr 3yr; do
kubectl exec "$POD" -- wget -qO- --post-data="test_routing{retention=\"${retention}\"} 1.0" \
"http://127.0.0.1:8429/api/v1/import/prometheus"
done
```
Query the data back from vmselect (it may take around 30-60 seconds for new data to be available for queries):
```shell
for retention in 3mo 1yr 3yr; do
echo "-> retention=${retention}"
curl -s "http://localhost:8481/select/0/prometheus/api/v1/query" \
--data-urlencode "query=test_routing{retention=\"${retention}\"}"
echo
done
```
You can also check that vmagent is forwarding data to all three groups:
```shell
curl -s http://localhost:8429/metrics | grep vmagent_remotewrite_blocks_sent_total
```
Each `url="N:secret-url"` corresponds to one `remoteWrite` entry (N=1 for Group A, N=2 for Group B, N=3 for Group C). Non-zero values confirm data is flowing.
## Alternative Routing by Existing Labels
The example setup above relies on a synthetic `retention` label to exist in every incoming metric.
If having a `retention` label in every metric isn't practical, you can, as an alternative, rely on existing labels to map data to the correct storage group.
The following example configures `vmagent` to route metrics based on the `environment` and `team` labels:
```yaml
# vmagent.yaml
remoteWrite:
# send dev and staging data to Group A
- url: "http://vmcluster-a-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write"
urlRelabelConfig:
- action: keep
source_labels: [environment]
regex: "dev|staging"
# send prod data to Group B
- url: "http://vmcluster-b-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write"
urlRelabelConfig:
- action: keep
source_labels: [environment]
regex: "prod|production"
# send data from Infra and SRE teams to Group B
- url: "http://vmcluster-b-victoria-metrics-cluster-vminsert:8480/insert/0/prometheus/api/v1/write"
urlRelabelConfig:
- action: keep
source_labels: [team]
regex: "infra|sre"
```
> Metrics that do not match any of the `keep` rules are dropped in the configuration above.
## Alternative Multi-Tenant Routing
VictoriaMetrics Cluster supports [multiple isolated tenants](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#multitenancy) identified by `accountID` (and optionally `projectID`) in the URL path, e.g., `/insert/0/prometheus/...`.
In a standard deployment, a single `vminsert` handles all tenants and distributes data across a shared pool of `vmstorage` nodes. In our current setup, each retention group deploys its own `vminsert`. This means tenant IDs are not required for routing, since data is already isolated by the URL that receives the write.
You can safely use a single tenant (`/insert/0/prometheus`) for all groups and rely on the `retention` label or label-based routing to separate data at query time.
If, however, you prefer tenant-level separation for the query layer, you can assign each group a distinct tenant ID, for instance:
| Group | Insert URL | Query URL |
|-------|------------|-----------|
| A (3mo) | `/insert/0/prometheus` | `/select/0/prometheus` |
| B (1yr) | `/insert/1/prometheus` | `/select/1/prometheus` |
| C (3yr) | `/insert/2/prometheus` | `/select/2/prometheus` |
This lets you query a single retention group directly, e.g., `/select/1/prometheus/api/v1/query?query=up` returns only data written to group B's `vminsert`. Queries to vmselect without a tenant prefix aggregate across all groups, preserving the unified view for dashboards.
The tenant path does not affect where data is stored (routing is always determined by which `vminsert` receives the data). The tenant ID is purely a query-scoping convenience in this architecture.
## Additional Enhancements
You can set up [vmauth](https://docs.victoriametrics.com/victoriametrics/vmauth/) to route data to the specified vminsert group based on the required retention.
Every group of vmstorages can handle one tenant or multiple one. Different groups can have overlapping tenants. As vmselect reads from all vmstorage nodes, the data is aggregated on its level.
**Additional Enhancements**
You can set up [vmauth](https://docs.victoriametrics.com/victoriametrics/vmauth/) for routing data to the given vminsert group depending on the needed retention.

View File

@@ -205,13 +205,15 @@ curl 'http://vmselect:8481/select/multitenant/prometheus/api/v1/query' \
The precedence for applying filters for tenants follows this order:
1. Filter tenants by `extra_label` and `extra_filters` filters.
1. Filter tenants by `extra_label`, `extra_filters` and `extra_filters[]` filters.
These filters have the highest priority and are applied first when provided through the query arguments.
Filters use `OR` logic - a tenant is selected if it matches any of the filters.
2. Filter tenants from labels selectors defined at metricsQL query expression.
> **Security considerations**
It is recommended restricting access to `multitenant` endpoints only to trusted sources,
since untrusted source may break per-tenant data by writing unwanted samples or get access to data of arbitrary tenants.
See also [vmauth security doc](https://docs.victoriametrics.com/victoriametrics/vmauth/#security).
## Binaries

View File

@@ -1136,6 +1136,8 @@ By default, the last point on the interval `[now - max_lookback ... now]` is scr
For instance, `/federate?match[]=up&max_lookback=1h` would return last points on the `[now - 1h ... now]` interval. This may be useful for time series federation
with scrape intervals exceeding `5m`.
VictoriaMetrics supports Prometheus v3.0 utf-8 content encoding with `Accept` header. If `Accept: allow-utf-8` HTTP header provided, `/federate` API response changes according to [Prometheus utf-8](https://prometheus.io/docs/guides/utf8/#querying) specification - `metric_name{tag="value"}` transforms into `{"metric_name","tag"="value"}`.
## Capacity planning
VictoriaMetrics uses lower amounts of CPU, RAM and storage space on production workloads compared to competing solutions (Prometheus, Thanos, Cortex, TimescaleDB, InfluxDB, QuestDB, M3DB) according to [our case studies](https://docs.victoriametrics.com/victoriametrics/casestudies/).

View File

@@ -26,10 +26,29 @@ See also [LTS releases](https://docs.victoriametrics.com/victoriametrics/lts-rel
## tip
* FEATURE: [vmagent](https://docs.victoriametrics.com/victoriametrics/vmagent/), [vmsingle](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) and `vminsert` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): add `-opentelemetry.promoteAllResourceAttributes` and `-opentelemetry.promoteScopeMetadata` command-line flags to allow managing label promotion for resource attributes and OTel scope metadata. See [OpenTelemetry](https://docs.victoriametrics.com/victoriametrics/integrations/opentelemetry/) docs and [#10931](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10931).
## [v1.145.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v1.145.0)
Release candidate
* SECURITY: upgrade Go builder from Go1.26.3 to Go1.26.4. See [the list of issues addressed in Go1.26.4](https://github.com/golang/go/issues?q=milestone%3AGo1.26.4%20label%3ACherryPickApproved).
* FEATURE: [enterprise](https://docs.victoriametrics.com/enterprise/) [vmsingle](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) and `vmstorage` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): add the new metrics `vm_downsampling_partitions_scheduled_rows` and `vm_retention_filters_partitions_scheduled_rows` for measuring background historical data merge completion time. See [#10960](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10960)
* FEATURE: [vmalert](https://docs.victoriametrics.com/victoriametrics/vmalert/): support `match[]=<label_selector>` query parameters in `/api/v1/rules` and `/api/v1/alerts` APIs to return only the rules that have configured labels satisfying the provided label selectors. See [11020](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11020).
* FEATURE: [vmagent](https://docs.victoriametrics.com/victoriametrics/vmagent/), [vmsingle](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) and `vminsert` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): add `-opentelemetry.promoteAllResourceAttributes` and `-opentelemetry.promoteScopeMetadata` command-line flags to allow managing label promotion for resource attributes and OTel scope metadata. See [OpenTelemetry](https://docs.victoriametrics.com/victoriametrics/integrations/opentelemetry/) docs and [#10931](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10931).
* FEATURE: [vmagent](https://docs.victoriametrics.com/vmagent/) : introduce `vmagent_remotewrite_kafka_outbuf_latency_seconds` and `vmagent_remotewrite_kafka_rtt_seconds` metrics for [kafka integration](https://docs.victoriametrics.com/victoriametrics/integrations/kafka/). The metrics could help identify throughput bottlenecks. See [#10730](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10730).
* FEATURE: [vmauth](https://docs.victoriametrics.com/victoriametrics/vmauth/): properly log user information when a missing route error occurs. See [#11052](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11052).
* FEATURE: [vmctl](https://docs.victoriametrics.com/vmctl/): add the ability to migrate data from [Mimir](https://docs.victoriametrics.com/victoriametrics/vmctl/mimir/#) object storage to VictoriaMetrics. See [#7717](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/7717).
* FEATURE: [dashboards](https://github.com/VictoriaMetrics/VictoriaMetrics/tree/master/dashboards): show the full `version` label in the `Version` panel when `short_version` label is empty (e.g. custom builds from feature branch). Previously, the panel could appear empty. See [#11047](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11047).
* BUGFIX: [vmalert](https://docs.victoriametrics.com/victoriametrics/vmalert/): fix the `Notifiers` page in web UI appearing blank despite the API returning notifier data correctly. See [#11035](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11035).
* BUGFIX: [vmalert](https://docs.victoriametrics.com/victoriametrics/vmalert/): reset the group evaluation timestamp if it exceeds the current host time. Previously, vmalert could use future timestamps for evaluations if the system clock was shifted backward. See [#10985](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10985).
* BUGFIX: [vmagent](https://docs.victoriametrics.com/victoriametrics/vmagent/) and [vmsingle](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/): properly parse [Prometheus Native Histograms](https://prometheus.io/docs/specs/native_histograms/), previously Protobuf parser could produce unexpected `vmrange` labels. See [#11041](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11041).
* BUGFIX: [vmauth](https://docs.victoriametrics.com/victoriametrics/vmauth/): properly calculate number of loaded users to be printed in startup log. Previously, it was only accounting for static users and skipped JWT configuration entries. See [#11050](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11050/).
* BUGFIX: [MetricsQL](https://docs.victoriametrics.com/victoriametrics/metricsql/): `integrate()` no longer extrapolates the last sample's value past the end of the time series. Previously, querying `integrate(metric[1h])` at a timestamp where the series had already ended would keep accruing area as if the last value continued indefinitely, producing values much larger than the true integral. See [#9474](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/9474). Thanks to @wtfashwin for contribution.
* BUGFIX: `vmselect` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): avoid returning HTTP 503 for queries with partial results when a storage group is unavailable and `-search.denyPartialResponse` is disabled. See [#11009](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/11009). Thanks to @fxrlv for the contribution.
* BUGFIX: [vmsingle](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) and `vmselect` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): properly escape `utf-8` label names for [/federate](https://docs.victoriametrics.com/victoriametrics/#federation) API requests. See [#10968](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10968).
* BUGFIX: [vmui](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/#vmui): persist the `Disable deduplication` toggle under its own local storage key. Before this fix, the toggle state was lost after reload and could overwrite the `Compact view` table setting. See [#11004](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/11004). Thanks to @immanuwell for the contribution.
* BUGFIX: `vmselect` in [VictoriaMetrics cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/): fix intermittent `write: connection timed out` errors caused by silently dropped TCP connections being reused from the connection pool. See [#10735](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/10735#issuecomment-4535832301).
## [v1.144.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v1.144.0)

View File

@@ -502,7 +502,7 @@ See the docs at https://docs.victoriametrics.com/victoriametrics/vmagent/ .
Supports array of values separated by comma or specified via multiple flags.
Empty values are set to default value.
-remoteWrite.roundDigits array
Round metric values to this number of decimal digits after the point before writing them to remote storage. Examples: -remoteWrite.roundDigits=2 would round 1.236 to 1.24, while -remoteWrite.roundDigits=-1 would round 126.78 to 130. By default, digits rounding is disabled. Set it to 100 for disabling it for a particular remote storage. This option may be used for improving data compression for the stored metrics (default 100)
Round metric values to this number of decimal digits after the point before writing them to remote storage. Examples: -remoteWrite.roundDigits=2 would round 1.236 to 1.24, while -remoteWrite.roundDigits=-1 would round 126.78 to 130. By default, digits rounding is disabled. Set it to 100 for disabling it for a particular remote storage. This option may be used for improving data compression for the stored metrics. See also -remoteWrite.significantFigures (default 100)
Supports array of values separated by comma or specified via multiple flags.
Empty values are set to default value.
-remoteWrite.sendTimeout array

View File

@@ -801,17 +801,15 @@ Please refer to the [VictoriaMetrics Cloud documentation](https://docs.victoriam
`vmalert` runs a web-server (`-httpListenAddr`) for serving metrics and alerts endpoints:
* `http://<vmalert-addr>` - UI;
* `http://<vmalert-addr>/api/v1/rules` - list of all loaded groups and rules. Supports `search`, `group_limit`, and `page_num` parameters, as well as additional [filtering](https://prometheus.io/docs/prometheus/latest/querying/api/#rules);
* `http://<vmalert-addr>/api/v1/alerts` - list of all active alerts;
* `http://<vmalert-addr>/api/v1/notifiers` - list all available notifiers;
* `http://<vmalert-addr>/vmalert/api/v1/alert?group_id=<group_id>&alert_id=<alert_id>` - get alert status in JSON format.
* `http://<vmalert-addr>/vmalert/api/v1/rule?group_id=<group_id>&rule_id=<rule_id>` - get rule status in JSON format.
* `http://<vmalert-addr>/vmalert/api/v1/group?group_id=<group_id>` - get group status in JSON format.
Used as alert source in AlertManager.
* `http://<vmalert-addr>/vmalert/alert?group_id=<group_id>&alert_id=<alert_id>` - get alert status in web UI.
* `http://<vmalert-addr>/vmalert/rule?group_id=<group_id>&rule_id=<rule_id>` - get rule status in web UI.
* `http://<vmalert-addr>/vmalert/api/v1/rule?group_id=<group_id>&alert_id=<alert_id>` - get rule status in JSON format.
* `http://<vmalert-addr>/metrics` - application metrics.
* `http://<vmalert-addr>/api/v1/rules` - returns a list of all loaded groups and rules. Supports the `datasource_type`, `search`, `group_limit`, and `page_num` parameters, as well as additional [filtering](https://prometheus.io/docs/prometheus/latest/querying/api/#rules);
* `http://<vmalert-addr>/api/v1/alerts` - returns a list of all active alerts. Supports the `datasource_type`, `rule_group[]`, `file[]` and `match[]`(applied on templated alert labels) query parameters;
* `http://<vmalert-addr>/api/v1/notifiers` - returns a list of all available notifiers;
* `http://<vmalert-addr>/vmalert/api/v1/alert?group_id=<group_id>&alert_id=<alert_id>` - returns the alert status in JSON format;
* `http://<vmalert-addr>/vmalert/api/v1/rule?group_id=<group_id>&rule_id=<rule_id>` - returns the rule status in JSON format;
* `http://<vmalert-addr>/vmalert/api/v1/group?group_id=<group_id>` - returns the group status in JSON format. Used as the alert source in AlertManager;
* `http://<vmalert-addr>/vmalert/alert?group_id=<group_id>&alert_id=<alert_id>` - displays the alert status in the web UI;
* `http://<vmalert-addr>/vmalert/rule?group_id=<group_id>&rule_id=<rule_id>` - displays the rule status in the web UI;
* `http://<vmalert-addr>/metrics` - application metrics endpoint;
* `http://<vmalert-addr>/-/reload` - hot configuration reload.
`vmalert` web UI can be accessed from [single-node version of VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/)

View File

@@ -1536,6 +1536,16 @@ To enable TLS on the public listener while keeping the internal listener non-TLS
`vmauth` also supports restricting access by IP - see [these docs](#ip-filters). See also [concurrency limiting docs](#concurrency-limiting).
When `vmauth` performs tenant routing for [multitenant](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#multitenant-reads) requests, it is crucial to explicitly set `extra_label`, `extra_filters` and `extra_filters[]` in the url_prefix configuration:
```yaml
unauthorized_user:
url_prefix: http://vmselect/select/multitenant?extra_filters[]=&extra_filters=&extra_label=vm_account_id=10&extra_label=vm_project_id=100
```
This is required because `vmselect` uses `OR` logic for tenant filtering. If a client sets `extra_filters[]` or `extra_filters`, it could bypass the tenant restriction configured via `extra_label`.
## Automatic issuing of TLS certificates
`vmauth` [Enterprise](https://docs.victoriametrics.com/victoriametrics/enterprise/) supports automatic issuing of TLS certificates via [Let's Encrypt service](https://letsencrypt.org/).

View File

@@ -73,4 +73,66 @@ ou can define it via the flag `--remote-read-headers=X-Scope-OrgID:demo`.
See [remote-read mode](https://docs.victoriametrics.com/victoriametrics/vmctl/remoteread/) for more details.
See also general [vmctl migration tips](https://docs.victoriametrics.com/victoriametrics/vmctl/#migration-tips).
See also general [vmctl migration tips](https://docs.victoriametrics.com/victoriametrics/vmctl/#migration-tips).
### Read data from the remote storage like S3, GCS, Azure etc.
If you have data stored in remote storage like S3, GCS, Azure etc. you can use `vmctl` in `mimir` mode to read data from
the remote storage and import it into VictoriaMetrics. In this mode `vmctl` reads data from the remote storage or file system
and checks index file, define needed blocks to be processed. After it downloads blocks by defined filters and
use Prometheus converter to read and sent data to VictoriaMetrics.
The following example shows how to read data from the file system and import it into VictoriaMetrics:
```sh
./vmctl mimir --mimir-path="fs:///mimir/test_data/mimir-tsdb" \ ? ? orbstack
--mimir-tenant-id=anonymous \
--mimir-filter-time-start=2024-12-01T00:00:00 \
--mimir-filter-time-end=2024-12-18T23:59:59 \
--mimir-creds-file-path=creads \
--vm-concurrency=6 \
--mimir-concurrency=6 \
--vm-addr=http://localhost:8428/
```
This approach is useful when you have data stored on the local file system or you have a mounted volume,
download the data from the remote storage etc.
The following example shows how to read data from the remote storage and import it into VictoriaMetrics:
```sh
./vmctl mimir --mimir-path="s3:///mimir-tsdb/anonymous" \
--mimir-filter-time-start=2024-12-01T00:00:00 \
--mimir-filter-time-end=2024-12-17T23:59:59 \
--mimir-creds-file-path=creads \
--mimir-custom-s3-endpoint='http://localhost:9000' \
--vm-concurrency=6 \
--mimir-concurrency=6 \
--vm-addr=http://localhost:8428/
```
In the example above we are used `--mimir-custom-s3-endpoint` flag to specify the custom S3 endpoint if it is needed.
When the process finishes, you will see the following:
```sh
2025/01/18 13:01:59 Fetching blocks from remote storage
Found 204 blocks to import. Continue? [Y/n] y
VM worker 0:? 1589405 samples/s
VM worker 1:? 1911834 samples/s
VM worker 2:? 1849187 samples/s
VM worker 3:? 1648820 samples/s
VM worker 4:? 1539212 samples/s
VM worker 5:? 1411485 samples/s
Processing blocks: 204 / 204 [?????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????] 100.00%
2025/01/18 13:02:18 Import finished!
2025/01/18 13:02:18 VictoriaMetrics importer stats:
idle duration: 18.485875611s;
time spent while importing: 16.40543875s;
total samples: 177961995;
samples/s: 10847743.71;
total bytes: 4.1 GB;
bytes/s: 248.2 MB;
import requests: 893;
import requests retries: 0;
2025/01/18 13:02:18 Total time: 18.867547083s
```

View File

@@ -56,7 +56,7 @@ OPTIONS:
--vm-compress Whether to apply gzip compression to import requests (default: true)
--vm-batch-size value How many samples importer collects before sending the import request to VM (default: 200000)
--vm-significant-figures value The number of significant figures to leave in metric values before importing. See https://en.wikipedia.org/wiki/Significant_figures. Zero value saves all the significant figures. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-round-digits option (default: 0)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics (default: 100)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-significant-figures option (default: 100)
--vm-extra-label value [ --vm-extra-label value ] Extra labels, that will be added to imported timeseries. In case of collision, label value defined by flag will have priority. Flag can be set multiple times, to add few additional labels.
--vm-rate-limit value Optional data transfer rate limit in bytes per second.
By default, the rate limit is disabled. It can be useful for limiting load on configured via '--vm-addr' destination. (default: 0)

View File

@@ -51,7 +51,7 @@ OPTIONS:
--vm-compress Whether to apply gzip compression to import requests (default: true)
--vm-batch-size value How many samples importer collects before sending the import request to VM (default: 200000)
--vm-significant-figures value The number of significant figures to leave in metric values before importing. See https://en.wikipedia.org/wiki/Significant_figures. Zero value saves all the significant figures. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-round-digits option (default: 0)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics (default: 100)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-significant-figures option (default: 100)
--vm-extra-label value [ --vm-extra-label value ] Extra labels, that will be added to imported timeseries. In case of collision, label value defined by flag will have priority. Flag can be set multiple times, to add few additional labels.
--vm-rate-limit value Optional data transfer rate limit in bytes per second.
By default, the rate limit is disabled. It can be useful for limiting load on configured via '--vm-addr' destination. (default: 0)

View File

@@ -44,7 +44,7 @@ OPTIONS:
--vm-compress Whether to apply gzip compression to import requests (default: true)
--vm-batch-size value How many samples importer collects before sending the import request to VM (default: 200000)
--vm-significant-figures value The number of significant figures to leave in metric values before importing. See https://en.wikipedia.org/wiki/Significant_figures. Zero value saves all the significant figures. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-round-digits option (default: 0)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics (default: 100)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-significant-figures option (default: 100)
--vm-extra-label value [ --vm-extra-label value ] Extra labels, that will be added to imported timeseries. In case of collision, label value defined by flag will have priority. Flag can be set multiple times, to add few additional labels.
--vm-rate-limit value Optional data transfer rate limit in bytes per second.
By default, the rate limit is disabled. It can be useful for limiting load on configured via '--vm-addr' destination. (default: 0)

View File

@@ -59,7 +59,7 @@ OPTIONS:
--vm-compress Whether to apply gzip compression to import requests (default: true)
--vm-batch-size value How many samples importer collects before sending the import request to VM (default: 200000)
--vm-significant-figures value The number of significant figures to leave in metric values before importing. See https://en.wikipedia.org/wiki/Significant_figures. Zero value saves all the significant figures. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-round-digits option (default: 0)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics (default: 100)
--vm-round-digits value Round metric values to the given number of decimal digits after the point. This option may be used for increasing on-disk compression level for the stored metrics. See also --vm-significant-figures option (default: 100)
--vm-extra-label value [ --vm-extra-label value ] Extra labels, that will be added to imported timeseries. In case of collision, label value defined by flag will have priority. Flag can be set multiple times, to add few additional labels.
--vm-rate-limit value Optional data transfer rate limit in bytes per second.
By default, the rate limit is disabled. It can be useful for limiting load on configured via '--vm-addr' destination. (default: 0)

6
go.mod
View File

@@ -1,6 +1,6 @@
module github.com/VictoriaMetrics/VictoriaMetrics
go 1.26.3
go 1.26.4
require (
cloud.google.com/go/storage v1.62.1
@@ -25,6 +25,7 @@ require (
github.com/googleapis/gax-go/v2 v2.22.0
github.com/influxdata/influxdb v1.12.4
github.com/klauspost/compress v1.18.5
github.com/oklog/ulid/v2 v2.1.1
github.com/prometheus/prometheus v0.311.3
github.com/urfave/cli/v2 v2.27.7
github.com/valyala/fastjson v1.6.10
@@ -109,7 +110,6 @@ require (
github.com/modern-go/reflect2 v1.0.3-0.20250322232337-35a7c28c31ee // indirect
github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
github.com/mwitkow/go-conntrack v0.0.0-20190716064945-2f068394615f // indirect
github.com/oklog/ulid/v2 v2.1.1 // indirect
github.com/open-telemetry/opentelemetry-collector-contrib/internal/exp/metrics v0.150.0 // indirect
github.com/open-telemetry/opentelemetry-collector-contrib/pkg/pdatautil v0.150.0 // indirect
github.com/open-telemetry/opentelemetry-collector-contrib/processor/deltatocumulativeprocessor v0.150.0 // indirect
@@ -155,7 +155,7 @@ require (
go.uber.org/zap v1.27.1 // indirect
go.yaml.in/yaml/v2 v2.4.4 // indirect
go.yaml.in/yaml/v3 v3.0.4 // indirect
golang.org/x/crypto v0.51.0 // indirect
golang.org/x/crypto v0.52.0 // indirect
golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f // indirect
golang.org/x/sync v0.20.0 // indirect
golang.org/x/term v0.43.0 // indirect

4
go.sum
View File

@@ -509,8 +509,8 @@ go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
golang.org/x/crypto v0.51.0 h1:IBPXwPfKxY7cWQZ38ZCIRPI50YLeevDLlLnyC5wRGTI=
golang.org/x/crypto v0.51.0/go.mod h1:8AdwkbraGNABw2kOX6YFPs3WM22XqI4EXEd8g+x7Oc8=
golang.org/x/crypto v0.52.0 h1:RMs7fP2rXdep0CftQlK8Uf+kibLm7qkCcradZWYz988=
golang.org/x/crypto v0.52.0/go.mod h1:1QgfPxDqh0T2M/elOJtp9RvuR95kVjir0e6/BvEmGbc=
golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f h1:W3F4c+6OLc6H2lb//N1q4WpJkhzJCK5J6kUi1NTVXfM=
golang.org/x/exp v0.0.0-20260410095643-746e56fc9e2f/go.mod h1:J1xhfL/vlindoeF/aINzNzt2Bket5bjo9sdOYzOsU80=
golang.org/x/mod v0.2.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=

View File

@@ -398,6 +398,10 @@ func (nhctx *nativeHistogramContext) appendTimeSeries(tss []TimeSeries, baseLabe
if baseName == "" {
return tss, labelsPool, samplesPool
}
originName := *nameValueP
defer func() {
*nameValueP = originName
}()
*nameValueP = fb.formatName(baseName, "_count")
tss, labelsPool, samplesPool = appendHistogramSeries(tss, labelsPool, samplesPool, baseLabels, "", tsMillis, count)
@@ -550,9 +554,11 @@ func (nhctx *nativeHistogramContext) reset() {
nhctx.zeroCountInt = 0
nhctx.zeroCountFloat = 0
nhctx.timestamp = 0
clear(nhctx.negativeSpans)
nhctx.negativeSpans = nhctx.negativeSpans[:0]
nhctx.negativeDeltas = nhctx.negativeDeltas[:0]
nhctx.negativeCounts = nhctx.negativeCounts[:0]
clear(nhctx.positiveSpans)
nhctx.positiveSpans = nhctx.positiveSpans[:0]
nhctx.positiveDeltas = nhctx.positiveDeltas[:0]
nhctx.positiveCounts = nhctx.positiveCounts[:0]

View File

@@ -3,8 +3,9 @@ package prompb
import (
"encoding/binary"
"math"
"reflect"
"testing"
"github.com/google/go-cmp/cmp"
)
func TestUnmarshalTimeSeries(t *testing.T) {
@@ -18,8 +19,8 @@ func TestUnmarshalTimeSeries(t *testing.T) {
if err != nil {
t.Fatalf("unexpected error: %s", err)
}
if !reflect.DeepEqual(tss, wantTSS) {
t.Fatalf("unexpected result\ngot:\n%v\nwant:\n%v", tss, wantTSS)
if diff := cmp.Diff(wantTSS, tss); len(diff) > 0 {
t.Fatalf("unexpected timeseries (-want, +got):\n%s", diff)
}
}
@@ -197,6 +198,93 @@ func TestUnmarshalTimeSeries(t *testing.T) {
},
})
}
{
// verify histogram fields are correctly reused
nativeHistogramC := nativeHistogramContext{
countInt: 0,
isCountFloat: true,
countFloat: 2.5,
sum: 1.0,
schema: 1,
zeroThreshold: 0.00001,
isZeroCountFloat: true,
zeroCountFloat: 0.5,
timestamp: 3000,
positiveSpans: []bucketSpan{{offset: 1, length: 2}},
positiveCounts: []float64{1.5, 1.0},
negativeSpans: []bucketSpan{{offset: 0, length: 1}},
}
nativeHistogramC2 := nativeHistogramContext{
countInt: 0,
isCountFloat: true,
countFloat: 0,
sum: 1.0,
schema: 1,
zeroThreshold: 0.00001,
isZeroCountFloat: true,
zeroCountFloat: 0.5,
timestamp: 4000,
positiveSpans: []bucketSpan{{offset: 0, length: 2}},
positiveCounts: []float64{1.5, 1.0},
negativeSpans: []bucketSpan{{offset: 0, length: 1}},
negativeCounts: []float64{1.5, 0},
}
hd1 := encodeHistogram(nativeHistogramC)
hd2 := encodeHistogram(nativeHistogramC2)
src := encodeTimeSeries(
[]Label{{Name: "__name__", Value: "rpc_latency_seconds"}},
nil,
[][]byte{hd1, hd2},
)
f(src, []TimeSeries{
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_count"}},
Samples: []Sample{{Value: 2.5, Timestamp: 3000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_sum"}},
Samples: []Sample{{Value: 1.0, Timestamp: 3000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(-0.00001, 0.00001)}},
Samples: []Sample{{Value: 0.5, Timestamp: 3000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(1, 1.414)}},
Samples: []Sample{{Value: 1.5, Timestamp: 3000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(1.414, 2.0)}},
Samples: []Sample{{Value: 1.0, Timestamp: 3000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_count"}},
Samples: []Sample{{Value: 0, Timestamp: 4000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_sum"}},
Samples: []Sample{{Value: 1.0, Timestamp: 4000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(-0.00001, 0.00001)}},
Samples: []Sample{{Value: 0.5, Timestamp: 4000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(0.7071, 1)}},
Samples: []Sample{{Value: 1.5, Timestamp: 4000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(1, 1.414)}},
Samples: []Sample{{Value: 1.0, Timestamp: 4000}},
},
{
Labels: []Label{{Name: "__name__", Value: "rpc_latency_seconds_bucket"}, {Name: "vmrange", Value: appendVmrangeHelper(-1, -0.7071)}},
Samples: []Sample{{Value: 1.5, Timestamp: 4000}},
},
})
}
}
func encodeTimeSeries(labels []Label, samples []Sample, histograms [][]byte) []byte {

View File

@@ -20,7 +20,7 @@ func chacha20Poly1305Open(dst []byte, key []uint32, src, ad []byte) bool
func chacha20Poly1305Seal(dst []byte, key []uint32, src, ad []byte)
var (
useAVX2 = cpu.X86.HasAVX2 && cpu.X86.HasBMI2
useAVX2 = cpu.X86.HasSSSE3 && cpu.X86.HasAVX2 && cpu.X86.HasBMI2
)
// setupState writes a ChaCha20 input matrix to state. See
@@ -47,7 +47,7 @@ func setupState(state *[16]uint32, key *[32]byte, nonce []byte) {
}
func (c *chacha20poly1305) seal(dst, nonce, plaintext, additionalData []byte) []byte {
if !cpu.X86.HasSSSE3 {
if !useAVX2 {
return c.sealGeneric(dst, nonce, plaintext, additionalData)
}
@@ -66,7 +66,7 @@ func (c *chacha20poly1305) seal(dst, nonce, plaintext, additionalData []byte) []
}
func (c *chacha20poly1305) open(dst, nonce, ciphertext, additionalData []byte) ([]byte, error) {
if !cpu.X86.HasSSSE3 {
if !useAVX2 {
return c.openGeneric(dst, nonce, ciphertext, additionalData)
}

File diff suppressed because it is too large Load Diff

2
vendor/modules.txt vendored
View File

@@ -851,7 +851,7 @@ go.yaml.in/yaml/v2
# go.yaml.in/yaml/v3 v3.0.4
## explicit; go 1.16
go.yaml.in/yaml/v3
# golang.org/x/crypto v0.51.0
# golang.org/x/crypto v0.52.0
## explicit; go 1.25.0
golang.org/x/crypto/chacha20
golang.org/x/crypto/chacha20poly1305