In older versions of shinyproxy containers that failed to start existed for a little while longer so I could check their logs using docker logs <my-container>.
In version 3.2.2 they seem to get removed pretty quickly after exiting. Is this configurable? Can I somehow make them exist a little bit longer?
This solved the issue:
proxy:
...
container-log-path: ./container-logs
It saves logs of the individual containers.
Just to give a bit of context: this has indeed changed in version 3.2.0. The main advantage is that a user is immediately informed when a container fails to start up. Previously, the user would have to wait until a timeout was reached.
The container-log-path option is indeed a great solution! You could also consider using the ShinyProxy Operator , which includes a complete monitoring stack, allowing you to browse through logs and metrics.