DPE-3547 mitigations for container restart

[paulomach]

Issue

Workload containers get restarted due timeout on livenessProbe pebble endpoint. Discussion at juju lp bug: https://bugs.launchpad.net/bugs/2052517

Solution

Here some mitigations and optimizations, not a final solution though.

• set floor for max_connections in 100
• function retries
• flush logs in single call
• • test coverage