## Summary Roll out [the feature](https://gitlab.com/gitlab-org/gitlab/-/work_items/628132) currently behind the `mergeability_check_redis_marker` feature flag. - Introduced by: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/255070 - DRI: @marc_shaw - Team Slack channel: `#g_code_review` > [!note] > Process and guidance live in the docs - this issue is just the commands and a place to track the rollout. > "Rolling out" means incrementally enabling the flag on GitLab.com to validate stability - it is not the same as releasing the feature, which happens when the flag is removed. > [Feature flag controls](https://docs.gitlab.com/development/feature_flags/controls/) · [Feature flag lifecycle](https://handbook.gitlab.com/handbook/product-development/how-we-work/product-development-flow/feature-flag-lifecycle/#feature-flag-lifecycle) ## What could go wrong? Scope: `MergeRequests::MergeabilityCheckService#async_execute`, called from GET requests (merge request widget polls `cached_widget.json` and `widget.json`, the merge request `show` page, REST `GET .../merge_requests/:iid`). Flag off: writes `UPDATE merge_requests SET merge_status = 'checking'` inside the GET before enqueueing `MergeRequestMergeabilityCheckWorker`. Flag on: claims Redis key `merge_request:mergeability_check:<id>` (SET NX, 5 minute TTL) in `Gitlab::Redis::SharedState` and enqueues only if the claim wins; no database write in the GET. The worker shortens the TTL to 2 minutes when it starts and deletes the key when it finishes. `MergeRequest#checking?` returns true while the key exists, so REST `merge_status`, GraphQL `detailedMergeStatus`, and webhooks report `checking` as before. The widget's `widget.json` also stops running the mergeability check synchronously and reports `mergeable: false` until the worker writes the verdict; the widget shows the checking spinner and updates from the `mergeRequestMergeStatusUpdated` GraphQL subscription or its next poll. Blast radius: merge status display for open merge requests whose status was reset to `unchecked` (typically by a push to the target branch). Correctness of the final verdict is unchanged - it is still computed by the worker under the existing exclusive lease and stale-input check. Three risks: (1) the verdict arrives later since the first widget poll no longer computes it in-request; the wait is the worker's queue time (`urgency :low`, 1 minute SLO) plus delivery via subscription or poll; (2) a lost job leaves the key until its TTL expires, so a re-check can be delayed by up to 2 minutes (or 5 if the job never started); (3) flag cache skew (web on, Sidekiq off for up to a minute at flip time) causes the worker to not delete the key, making a following re-check wait for the TTL. No data written to the database changes shape; no schema change; rollback is immediate since flag off restores the old write path and stale keys expire on their own. Watch: the `detect_writes_on_get` log events for the merge request widget endpoints (should drop to zero as the flag rolls out), `MergeRequestMergeabilityCheckWorker` queue latency and duration on the Sidekiq worker detail dashboard (https://dashboards.gitlab.net/d/sidekiq-worker-detail/sidekiq3a-worker-detail?orgId=1&var-worker=MergeRequestMergeabilityCheckWorker), the rate of "Failed to check mergeability of merge request" errors in Kibana (a rise indicates duplicate jobs losing the lease, which points at the 2 minute running TTL being too short for the worker duration distribution), and user reports of the widget stuck on "Checking if merge request can be merged". Spot check: after a push to the target branch of an open merge request in gitlab-org/gitlab, the widget shows the spinner and then the merge button within a few seconds, and `SELECT merge_status` for that merge request never reads `checking`. ## Events - [Exceptions with mergeability_check_redis_marker:1](https://log.gprd.gitlab.net/app/discover#/?_g=(time:(from:now-1d,to:now))&_a=(index:'7092c4e2-4eb5-46f2-8305-a7da2edad090',query:(language:kuery,query:'json.exception.feature_flag_states.keyword:%22mergeability_check_redis_marker:1%22'),columns:!(json.exception.class,json.exception.message,json.extra.workflow_id,json.meta.caller_id),sort:!(!(json.time,desc)))) - [Events with mergeability_check_redis_marker:1](https://log.gprd.gitlab.net/app/discover#/?_g=(time:(from:now-1d,to:now))&_a=(index:'7092c4e2-4eb5-46f2-8305-a7da2edad090',query:(language:kuery,query:'json.feature_flag_states:%22mergeability_check_redis_marker:1%22'),columns:!(json.meta.caller_id,json.feature_flag_states),sort:!(!(json.time,desc)))) - [Error rate and other graphs by modifying the examples in the Visualization Library](https://log.gprd.gitlab.net/app/visualize#/?_g=h@358d019&s=FF%20Observability) Feature Flag events are only logged by default for feature flags marked for the current or future milestones. To enable while the feature flag is active, see https://docs.gitlab.com/development/feature_flags/#logging ## Rollout Run all production `/chatops` in [`#production`](https://gitlab.slack.com/archives/C101F3796) and cross-post the results to `#g_code_review`. Background: [incremental rollout process](https://docs.gitlab.com/development/feature_flags/controls/#process), [feature actors](https://docs.gitlab.com/development/feature_flags/#feature-actors). **Non-production** ``` /chatops gitlab run feature set mergeability_check_redis_marker 50 --actors --dev --pre --staging --staging-ref /chatops gitlab run feature set mergeability_check_redis_marker true --dev --pre --staging --staging-ref ``` **Production** - start with `gitlab-org/gitlab`, then percentage of actors (wait at least 15 min between steps, watch the dashboards): ``` /chatops gitlab run feature set --project=gitlab-org/gitlab mergeability_check_redis_marker true /chatops gitlab run feature set mergeability_check_redis_marker <percentage> --actors ``` ## Before global rollout Confirm the relevant gotchas before going to 100% - see [enabling a feature for GitLab.com](https://docs.gitlab.com/development/feature_flags/controls/#enabling-a-feature-for-gitlabcom): - [Docs + version history](https://docs.gitlab.com/development/documentation/feature_flags/) updated - [Breaking changes](https://docs.gitlab.com/development/documentation/release_notes/#deprecations-removals-and-breaking-changes) announced, if any - [Change management issue](https://handbook.gitlab.com/handbook/engineering/infrastructure-platforms/change-management/#feature-flags-and-the-change-management-process) opened, if required - [External API consumers](https://docs.gitlab.com/development/feature_flags/#do-not-use-feature-flags-in-external-api-consumers) handled with a fail-open mechanism, if applicable ## Cleanup Remove the flag once [deemed stable](https://handbook.gitlab.com/handbook/product-development/how-we-work/product-development-flow/feature-flag-lifecycle/#feature-flag-lifecycle) - see [cleaning up](https://docs.gitlab.com/development/feature_flags/controls/#cleaning-up). Track it here, or open a follow-up [Feature Flag Cleanup issue](https://gitlab.com/gitlab-org/gitlab/-/work_items/new?description_template=Feature%20Flag%20Cleanup). Remove the flag and its YAML definition from the codebase, then: ``` /chatops gitlab run release check <merge-request-url> <milestone> /chatops gitlab run feature delete mergeability_check_redis_marker --dev --pre --staging --staging-ref --production ``` ## Rollback ``` /chatops gitlab run feature set mergeability_check_redis_marker false # production /chatops gitlab run feature set mergeability_check_redis_marker false --dev --pre --staging --staging-ref # non-production /chatops gitlab run feature delete mergeability_check_redis_marker --dev --pre --staging --staging-ref --production # remove entirely ```