Failures and restart deploy

An agent session can fail during infrastructure startup, source analysis, configuration synchronization, or the build handoff.

Review the failure

1. Open the current environment.
2. Review the agent view or recorded error.
3. Check Build Logs if the managed build started.
4. Check detected root directory, port, commands, Dockerfile, and variables.
5. Correct repository or environment settings that belong to your application.

The session records an error message and failure phase when available. Temporary agent infrastructure is cleaned up after failure.

Restart deploy

The dashboard exposes restart-deploy behavior for supported failures in the pre-build or post-build path. Restarting uses the current environment and source configuration to make another attempt.

If the failure is owned by your repository—for example a missing dependency, application exception, or missing variable—fix the repository or environment configuration first.

If Kuberns Fixes provides a supported Kuberns-side Dockerfile change, you can apply it from the dashboard and trigger another deployment.