Ensure Collector is always cleaned up

[Stringy]

Description

Sometimes Collector is either expected to fail or can fail during test setup or healthchecks. We want to retain any logs from the collector container, even in these cases.

The clean up function has been changed to unconditionally clean up Collector:

• Collector is not run with --rm so we can get the logs. For this reason, all the collector manager functions cope with a container that is not running (which makes the check in the clean up function redundant)
• The base class should be more thorough to ensure that all resources are cleaned up

Checklist

• Investigated and inspected CI test results

Testing Performed

CI primarily. Not able to strictly reproduce the problem seen here but existence of logs in the right places should be enough