Skip to content
Dekglas
Logmarq documentation

Troubleshooting

Separate startup, authentication, provider and record problems while preserving data and collecting a small safe report.

Troubleshooting

Start with the failure you can observe and the version actually running. Avoid deleting the data volume or changing several integration settings at once: both can hide the original cause.

The application does not start

Run docker compose ps, inspect docker compose logs logmarq locally and request http://localhost:8080/api/health. Confirm the image digest, available disk space and permissions on the data volume. A migration failure requires the exact release's recovery guidance; do not edit database tables by hand. Keep the previous backup and failed instance for diagnosis.

Sign-in or a redirect fails

Check that the browser's origin matches LOGMARQ_PUBLIC_URL, the reverse proxy terminates TLS correctly and secure cookies are configured for HTTPS. Check the provider's registered callback against the current release's instructions. Do not disable authentication or expose an alternate unprotected port to bypass the problem. A valid sign-in does not grant a missing application role.

GitHub connects but events do not arrive

A localhost or private address may be unreachable from GitHub. Check the App's webhook URL, selected event types and configured shared secret without disclosing the secret. Confirm that the App is installed on the relevant repository. A connected App and a successful read are not proof of webhook delivery. A delivery refusal or missing permission is not evidence of no changes.

Private package access uses its own supported sign-in flow. If that flow reports unavailable, its OAuth configuration may be missing; adding permissions to an unrelated App does not fix it.

A release or comparison looks wrong

Inspect the recorded commit and artifact digests, source configuration and comparison direction. A tag can move, and a discovered tag alone does not establish that a version shipped. Review any import preview after changing its settings. Correct the record using the product's ordinary confirmation and history paths so the reason remains attributable.

A scan fails

Check disk capacity, the configured artifact source and access to vulnerability definitions. Scanning and its definition download are opt-in. A missing or failed scan is not a clean result; retain its actual status rather than publishing an unsupported security conclusion.

Prepare a safe report

Record the version and digest, deployment method, failing action, time, expected result, actual result and minimal reproduction. Inspect logs locally and remove credentials, headers, customer names, private repository content and raw release data before sharing a short excerpt. Never send the whole database, volume, environment file, activation key or provider token.

Use the support contact confirmed for your installation. Report suspected vulnerabilities through the private security contact. For upgrade failures, see upgrades and support.