System diagnostics answers "what is the running system built from, and what is it actually connected to?" It is the first screen to open right after installation, right after an upgrade, and whenever something looks wrong. Passwords, access keys and tokens are not shown, and addresses are shown only up to scheme://host:port.

System diagnostics

The screen has four parts.

PartContents
Overall verdict (총평)Healthy (정상) / Needs attention (주의 필요) / Down (장애), version, time of the last check and how long it took
Deployment info (배포 정보)Commit and build time of the frontend and the backend — check that the deployment really changed
Dependent components (의존 컴포넌트)Reachability and extra details of PostgreSQL, object storage (per bucket), Prometheus, Alertmanager, Kubernetes, MLflow tracking and so on. Two per-tenant cards (tenant namespaces, MLflow workspaces) are here too
Configuration hygiene (설정 위생)Verdicts on settings that cause problems in production. Only those needing action appear as cards; passed ones appear only as a count

Run a check

  1. Open the screen. On opening, it actually pokes every dependent service once.
  2. To check again, press Check again (다시 점검) at the top right. There is no automatic refresh (the check is a heavy call).
  3. Scroll down and review Configuration hygiene. Each card states what is wrong and which environment variable to fix.
    System diagnostics — configuration hygiene

The two per-tenant cards

These two cards check each active tenant rather than shared infrastructure. They are healthy when every tenant has the resource, down when none do, and a warning when some are missing. Expand a card to see the result per tenant.

CardWhat it checksWhat breaks without itHow to fix
Tenant namespaces (테넌트 네임스페이스)The cluster has an Active tenant-<code> namespaceTraining and serving submissions for that tenant are rejected. The Kubernetes card only looks at nodes, so it stays greenkubectl create namespace tenant-<code>. It is created with the tenant, but is missing if you rebuilt the cluster or deleted it by hand. If it says "deleting", wait for the tenant deletion to finish
MLflow workspaces (MLflow 워크스페이스)Each tenant has its MLflow workspaceTraining runs to the end, but metric logging and model registration fail with 404Restart the app; startup recreates every tenant's workspace. This typically happens after geo-mlops-admin db reset while the app was running

The screen above is a local test environment, so it also shows several WARNINGs. On a production server, also clean up the following.

  • If you run behind HTTPS, GEO_MLOPS_AUTH_COOKIE_SECURE=true
  • Turn off the development mock DataOps server — GEO_MLOPS_MOCK_DATAOPS_ENABLED=false
  • Shared secret for the Alertmanager webhook — GEO_MLOPS_ALERTMANAGER_WEBHOOK_TOKEN
  • For CORS, put the real origins in GEO_MLOPS_CORS_ALLOW_ORIGINS instead of a wildcard

Written for the platform as of 2026-09-21.

© Geo-MLOps