Increase logging on startup to ensure we see configuration issues

Description

Updates to shibboleth caused unexpected behaviour, which are believed to be related to the newly-added gke-lb-health container, but hard to diagnose due to lack of logging in certain circumstances.

Logging should be improved to see more of these errors.

Further details

Make sure logs are written in "default" conditions, be more verbose when e.g. objects are None or not comparable or not found or length 0.

Task list

Acceptance criteria

• Container always logs pod statuses on startup and list all pods
• Container logs when no containers are found in the pod
• Container logs when overall status changes, including all checked container names
• Container always logs failure error codes
• Log when container is started

Links/references