Troubleshooting
Start at the narrowest failing layer and move outward. Avoid reinstalling or rebuilding everything before you know which boundary is broken.
A service will not start
- Open the latest deployment and logs.
- Confirm source/image and expected start command.
- Confirm required environment variables exist.
- Check the configured exposed port.
- Check linked datastore health and project networking.
- Review restart policy and container state.
A domain does not route
- Confirm the service itself is healthy privately.
- Check DNS against the server's public ingress.
- Check domain configuration on the service.
- Inspect proxy mode and generated configuration.
- Validate the server and relevant networks.
A server fails provisioning
Use the server connectivity test first, then review provisioning status. Confirm SSH access, supported Linux environment, Docker/Dokku state, and proxy-mode assumptions.
A manifest will not apply
Run preview/sync-plan and inspect validation errors. Check whether the plan contains removals that require confirmation, then inspect manifest status and drift.
Recovery concerns
Do not attempt a destructive restore blindly. Verify the destination, identify the backup/PITR target, and use an isolated recovery drill when possible.
Report a bug
When opening a GitHub issue, include the RailDock release/commit, affected service type, sanitized logs, the action you attempted, and the smallest reproducible sequence. Never paste secrets.