docs/vmanomaly: fix outdated anchors and param names

docs/anomaly-detection/CHANGELOG.md

@@ -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).

docs/anomaly-detection/FAQ.md

@@ -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 ...

docs/anomaly-detection/components/README.md

@@ -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.

docs/anomaly-detection/components/reader.md

@@ -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>