[FR] Improved clean-up options

[cx-michael-kubiaczyk]

Would be good to have a better way for e2e to clean up after itself when things fail in Cx1.
Example:

1. scan is stuck and cancelled by e2e
2. cancellation fails due to some backend issue, disconnect etc
3. delete-project fails due to a still-running stuck scan

Possible solutions:

1. retry-on-fail for any test, with optional delay?
2. filters on the executable to allow only specific test types similar to the current engine-filter flag
3. end-of-test cleanup phase that reruns read,delete statements? or tracks failed delete operations for specific retry? (scan-cancellation success response doesn't guarantee scan cancellation)