Integration tests leak resources

[voelzmo]

How to categorize this issue?

/area dev-productivity
/kind bug
/priority 3
/platform aws

What happened:
I recently PRed a change which had failing integration tests: #424. As a result, we got leaking resources in our TM cluster, namely cluster, infrastructure and things associated with these objects. Not only is this a waste of resources, it also required cleaning up the leaked resources before the fix would pass another IT run, as the old resources were still picked up by the reconciliation loop in the new test runs, causing it to fail.

What you expected to happen:
Even failing tests should not leak resources. Leaked resources from previous test runs should not cause new test runs to fail.

Anything else we need to know?:
This seems to be also true for other extension-providers, such as Azure.

[timebertt]

Ref gardener/terraformer#57 (comment) and #325 (comment)

[voelzmo]

Thanks for cross-linking! When I talked to @ialidzhikov about this, he mentioned this has been mentioned before somewhere! ;)