## Summary This issue is to roll out [the semantic code search REST API](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/227817) on production, which is currently behind the \`semantic_code_search_rest_api\` feature flag. ## Owners - Most appropriate Slack channel to reach out to: \`#g_global_search\` - Best individual to reach out to: @tgao3701908 ## Expectations ### What are we expecting to happen? The \`GET /api/v4/projects/:id/(-/)search/semantic\` endpoint becomes available to all users on instances where semantic code search is configured and indexed. ### What can go wrong and how would we detect it? - Unexpected load on the semantic search infrastructure — monitor via Elasticsearch/pgvector dashboards - Rate limit too permissive — the \`semantic_search_rate_limit\` provides a safety valve - The flag can be disabled per-installation if scalability issues arise (\`Feature.disable(:semantic_code_search_rest_api)\`) ## Rollout Steps ### Rollout on non-production environments - [ ] Verify the flag is enabled by default on staging - [ ] Confirm the endpoint returns 200 for a provisioned project on staging ### Rollout on GitLab.com - [ ] The flag is \`default_enabled: true\` — no explicit enable step needed - [ ] Monitor for errors or unexpected load after the MR merges - [ ] Disable per-group if issues arise: \`Feature.disable(:semantic_code_search_rest_api, Group.find(id))\` ### Cleanup - [ ] Remove \`semantic_code_search_rest_api\` feature flag - [ ] Remove \`route_setting :lifecycle, :beta\` from the endpoint - [ ] Update API documentation to reflect GA status