Implement retry for failure codes

[Zoltan-LS]

Ideally we should have a good baseline snapshot.
In case the baseline snapshot gives ex 500 , it should retry after a few seconds lets say 3 times until we get proper 200 or something resposne codes/screenshot

This would avoid false positives like :

At least some of them.

[dransome]

Maybe it's valid to retry every HTTP status code fail a few times? (i.e. not only the baseline)

IMO we care about "is this broken" rather than "is there an intermittent issue" - because (IMO) an intermittent issue is unlikely to be caused by server package updates / configuration issues (more likely to be server resource shortages or similar, which is outside of scope for this tool).