chore(infra): remove dead tempo ingester config - #3806
manamana32321 wants to merge 3 commits into
Conversation
The agent reaches traces through Grafana MCP, which registers its tempo_* tools only after probing /api/mcp on the Tempo datasource. Tempo serves that endpoint from 2.9 but keeps it off by default. Drop structuredConfig while here: the tempo chart has no such key (it belongs to tempo-distributed), so the ingester tuning under it was never rendered. Confirmed by helm template, which emits an empty ingester block both before and after this change. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configuration
📒 Files selected for processing (1)
💤 Files with no reviewable changes (1)
Included review availability: This review used your included allowance. Your plan provides up to 1 included review per hour; 0 remain after this review. 📝 WalkthroughWalkthroughThe Tempo values configuration no longer sets explicit ingester values for maximum block size, maximum block duration, or complete block timeout. ChangesTempo ingester settings
Priority: ⬇️ Low Estimated code review effort: 1 (Trivial) | ~3 minutes Change: Other Merge Risk: ⚪ Minimal · up to The change is a small values cleanup that is reported to leave the rendered Tempo configuration unchanged. No actionable merge-blocking risk was found. 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
The local stack ran Tempo 2.8.2, which predates the MCP server, so the endpoint the agent probes could not exist there. Bump it to 2.9.0 and turn the same flag on, keeping local and production aligned for this feature. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…423e Enabling Tempo's own MCP server adds no tools to the Grafana MCP server. Verified by diffing tools/list against two Grafana instances: one with Tempo 2.8.2 and no mcp_server, one with Tempo 2.9.0 and mcp_server enabled. Both expose the same 62 tools. mcp-grafana reaches Tempo through the Grafana datasource proxy (/api/datasources/proxy/uid/<uid>/api/search), not through /api/mcp. Its --disable-proxied flag is now an alias for --disable-tempo, so the proxied path these settings targeted no longer exists. The structuredConfig removal stays. That key does not exist in the tempo chart, so helm template renders identically with and without it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Description
infra/k8s/monitoring/tempo/values.yaml의structuredConfig블록을 지웁니다.이 키는 우리가 쓰는
tempo차트에 없는 키입니다(tempo-distributed차트 것). 그래서 그 아래 적어둔 ingester 설정값들은 한 번도 Tempo에 전달된 적이 없습니다. 동작은 그대로이고 오해만 일으키던 설정을 치우는 변경입니다.이 PR은 처음에 Tempo 자체 MCP 서버를 켜는 작업으로 시작했는데, 켜도 효과가 없다는 걸 확인해서 그 부분은 되돌렸습니다.
Additional context
죽은 설정이라는 근거 — 차트를 렌더해보면 삭제 전과 후의 출력이 완전히 같습니다.
MCP 서버 설정을 되돌린 근거 — Grafana를 두 개 띄워 연결된 도구 목록을 비교했습니다.
/api/mcp응답/api/mcp는 분명히 켜졌는데 도구가 한 개도 늘지 않았습니다. Grafana MCP는 Tempo에 그 주소로 접근하지 않고 Grafana의 데이터소스 프록시를 거치기 때문입니다. 트레이스 검색·조회·메트릭 질의는 이 설정 없이 이미 다 됩니다.Before submitting the PR, please make sure you do the following
Closes TAS-3079
🤖 Generated with Claude Code