Announce that legacy metric names are deprecated, will be turned off by default in Synapse v1.71.0 and removed altogether in Synapse v1.73.0. (#14024)

Author: reivilibre

changelog.d/14024.removal

@@ -0,0 +1 @@
+Announce that legacy metric names are deprecated, will be turned off by default in Synapse v1.71.0 and removed altogether in Synapse v1.73.0. See the upgrade notes for more information.

docs/metrics-howto.md

@@ -135,6 +135,8 @@ Synapse 1.2 updates the Prometheus metrics to match the naming
 convention of the upstream `prometheus_client`. The old names are
 considered deprecated and will be removed in a future version of
 Synapse.
+**The old names will be disabled by default in Synapse v1.71.0 and removed
+altogether in Synapse v1.73.0.**
 
 | New Name                                                                     | Old Name                                                               |
 | ---------------------------------------------------------------------------- | ---------------------------------------------------------------------- |
@@ -146,6 +148,13 @@ Synapse.
 | synapse_federation_client_events_processed_total                             | synapse_federation_client_events_processed                             |
 | synapse_event_processing_loop_count_total                                    | synapse_event_processing_loop_count                                    |
 | synapse_event_processing_loop_room_count_total                               | synapse_event_processing_loop_room_count                               |
+| synapse_util_caches_cache_hits                                               | synapse_util_caches_cache:hits                                         |
+| synapse_util_caches_cache_size                                               | synapse_util_caches_cache:size                                         |
+| synapse_util_caches_cache_evicted_size                                       | synapse_util_caches_cache:evicted_size                                 |
+| synapse_util_caches_cache                                                    | synapse_util_caches_cache:total                                        |
+| synapse_util_caches_response_cache_size                                      | synapse_util_caches_response_cache:size                                |
+| synapse_util_caches_response_cache_hits                                      | synapse_util_caches_response_cache:hits                                |
+| synapse_util_caches_response_cache_evicted_size                              | synapse_util_caches_response_cache:evicted_size                        |
 | synapse_util_metrics_block_count_total                                       | synapse_util_metrics_block_count                                       |
 | synapse_util_metrics_block_time_seconds_total                                | synapse_util_metrics_block_time_seconds                                |
 | synapse_util_metrics_block_ru_utime_seconds_total                            | synapse_util_metrics_block_ru_utime_seconds                            |
@@ -261,7 +270,7 @@ Standard Metric Names
 
 As of synapse version 0.18.2, the format of the process-wide metrics has
 been changed to fit prometheus standard naming conventions. Additionally
-the units have been changed to seconds, from miliseconds.
+the units have been changed to seconds, from milliseconds.
 
 | New name                                 | Old name                          |
 | ---------------------------------------- | --------------------------------- |

docs/upgrade.md

@@ -100,6 +100,34 @@ vice versa.
 Once all workers are upgraded to v1.69 (or downgraded to v1.68), receipts
 replication will resume as normal.
 
+
+## Deprecation of legacy Prometheus metric names
+
+In current versions of Synapse, some Prometheus metrics are emitted under two different names,
+with one of the names being older but non-compliant with OpenMetrics and Prometheus conventions
+and one of the names being newer but compliant.
+
+Synapse v1.71.0 will turn the old metric names off *by default*.
+For administrators that still rely on them and have not had chance to update their
+uses of the metrics, it's possible to specify `enable_legacy_metrics: true` in
+the configuration to re-enable them temporarily.
+
+Synapse v1.73.0 will **remove legacy metric names altogether** and it will no longer
+be possible to re-enable them.
+
+The Grafana dashboard, Prometheus recording rules and Prometheus Consoles included
+in the `contrib` directory in the Synapse repository have been updated to no longer
+rely on the legacy names. These can be used on a current version of Synapse
+because current versions of Synapse emit both old and new names.
+
+You may need to update your alerting rules or any other rules that depend on
+the names of Prometheus metrics.
+If you want to test your changes before legacy names are disabled by default,
+you may specify `enable_legacy_metrics: false` in your homeserver configuration.
+
+A list of affected metrics is available on the [Metrics How-to page](https://matrix-org.github.io/synapse/v1.69/metrics-howto.html?highlight=metrics%20deprecated#renaming-of-metrics--deprecation-of-old-names-in-12).
+
+
 # Upgrading to v1.68.0
 
 Two changes announced in the upgrade notes for v1.67.0 have now landed in v1.68.0.

docs/usage/configuration/config_documentation.md

@@ -2436,6 +2436,31 @@ Example configuration:
 enable_metrics: true
 ```
 ---
+### `enable_legacy_metrics`
+
+Set to `true` to publish both legacy and non-legacy Prometheus metric names,
+or to `false` to only publish non-legacy Prometheus metric names.
+Defaults to `true`. Has no effect if `enable_metrics` is `false`.
+**In Synapse v1.71.0, this will default to `false` before being removed in Synapse v1.73.0.**
+
+Legacy metric names include:
+- metrics containing colons in the name, such as `synapse_util_caches_response_cache:hits`, because colons are supposed to be reserved for user-defined recording rules;
+- counters that don't end with the `_total` suffix, such as `synapse_federation_client_sent_edus`, therefore not adhering to the OpenMetrics standard.
+
+These legacy metric names are unconventional and not compliant with OpenMetrics standards.
+They are included for backwards compatibility.
+
+Example configuration:
+```yaml
+enable_legacy_metrics: false
+```
+
+See https://github.com/matrix-org/synapse/issues/11106 for context.
+
+*Since v1.67.0.*
+
+**Will be removed in v1.73.0.**
+---
 ### `sentry`
 
 Use this option to enable sentry integration. Provide the DSN assigned to you by sentry

synapse/config/metrics.py

@@ -43,32 +43,6 @@ class MetricsConfig(Config):
     def read_config(self, config: JsonDict, **kwargs: Any) -> None:
         self.enable_metrics = config.get("enable_metrics", False)
 
-        """
-        ### `enable_legacy_metrics` (experimental)
-
-        **Experimental: this option may be removed or have its behaviour
-        changed at any time, with no notice.**
-
-        Set to `true` to publish both legacy and non-legacy Prometheus metric names,
-        or to `false` to only publish non-legacy Prometheus metric names.
-        Defaults to `true`. Has no effect if `enable_metrics` is `false`.
-
-        Legacy metric names include:
-        - metrics containing colons in the name, such as `synapse_util_caches_response_cache:hits`, because colons are supposed to be reserved for user-defined recording rules;
-        - counters that don't end with the `_total` suffix, such as `synapse_federation_client_sent_edus`, therefore not adhering to the OpenMetrics standard.
-
-        These legacy metric names are unconventional and not compliant with OpenMetrics standards.
-        They are included for backwards compatibility.
-
-        Example configuration:
-        ```yaml
-        enable_legacy_metrics: false
-        ```
-
-        See https://github.com/matrix-org/synapse/issues/11106 for context.
-
-        *Since v1.67.0.*
-        """
         self.enable_legacy_metrics = config.get("enable_legacy_metrics", True)
 
         self.report_stats = config.get("report_stats", None)