Skip to main content

Organization troubleshooting

Triage order​

  1. Identify the exact instance, version, host and time window.
  2. Check whether the Engine and Scheduler processes or container are running.
  3. Check host capacity: disk, memory, inode use and port ownership.
  4. Read the instance's Engine, Scheduler and container logs.
  5. Test the Engine locally before diagnosing the reverse proxy.
  6. Test authentication, one read-only page and one small model request separately.

Do not use a production instance as a test target and do not delete locks, journals or backups merely to make startup continue.

CLI-managed diagnostics​

Where the host CLI provides it:

clawuno status
clawuno doctor --check
clawuno doctor --report

The report is diagnostic evidence, not an authorization to repair Level 2 or data-related findings automatically.

Reverse proxy symptoms​

If local access works but remote access fails, verify HTTPS origin forwarding, WebSocket/streaming behavior, request size and timeout limits, and cookie origin consistency. A proxy health check should not restart the Engine merely because a long operation delayed an HTTP response.

For a lost administrator password, follow Recover an administrator account. This requires host access and never requires editing authentication JSON by hand.

Escalation package​

Provide the exact version and platform, deployment type, sanitized configuration, relevant log interval, reproduction steps, and generated diagnostic report. Remove credentials, setup tokens and customer content before sharing.