15.6 Container Logs

Go To Top

Return to Logging and Observability

Containers should expose logs in a way that supports runtime inspection, troubleshooting, and Evidence collection. The team should not depend on transient container output that disappears before reviewers can inspect it.

Container logs should identify:

1. Container name.
2. Image name and tag, where useful.
3. Node or service Identity.
4. Startup command.
5. Configuration profile.
6. Node Lifecycle events.
7. DDS discovery or communication issues, where relevant.
8. Warnings and errors.
9. Shutdown result.

The team may capture container logs via container runtime commands, mounted log directories, standard output and standard error, or evidence-export scripts. The container documentation should identify the expected mechanism.

If a container runs a single Node, the container logs and the Node logs may overlap. The team should still preserve the distinction between the container as a packaging/runtime unit and the Node as an architectural runtime participant.