Functional tests are unstable

[Sudeep Bhandari]

The functional tests for this charm are highly unstable. The test runs fail about 70-80% of the time.
There are 4 test bundles that we're deploying:
  - jammy-yoga
  - focal-yoga
  - jammy-yoga-ovn
  - focal-yoga-ovn
The failures don't look related to any of the bundles, because sometimes the tests pass in all the bundles, and sometimes they fail in any of the bundles.
The failures are seen mostly in the following functions (see file https://git.launchpad.net/charm-cloudsupport/tree/tests/functional/tests/modules/test_cloudsupport.py):
 - test_20_create_instance
 - test_25_test_connectivity

Sometimes, the failures can be avoided by retrying these tests for a few minutes, but sometimes even after retrying for upto 6 minutes, test_25_test_connectivity keeps failing.
Failures in test_20_create_instance have been mostly solved by addressing the issues in errored instance creation(by deleting and recreating the instances).
Some failures are on add_test_image (https://git.launchpad.net/charm-cloudsupport/tree/tests/functional/tests/modules/setup.py) but they disappear completely if we retry a few times.

We need to create more robust tests that pass most of the time, by finding the root causes of the failures, especially on test_25_test_connectivity.

Detailed logs on failures - https://code.launchpad.net/~sudeephb/charm-cloudsupport/+git/charm-cloudsupport/+merge/429864