fix(release-automation): retry regression bot comment reads

[hdamker]

What type of PR is this?

bug
tests

What this PR does / why we need it:

The release automation regression runner currently treats a failed comments API read as a final failure after the caller workflow has already succeeded. One canary run hit a transient/non-JSON gh api response while fetching Release Issue comments, even though the expected snapshot-created bot reply was present.

This PR retries issue-comment reads before reporting an infrastructure failure. It keeps the existing bot reply title checks unchanged.

Which issue(s) this PR fixes:

N/A - follow-up to #336.

Special notes for reviewers:

Validation:

• python3 -m pytest release_automation/tests/test_regression_runner.py -q
• python3 -m pytest release_automation/tests -q
• Full ReleaseTest round trip from this branch: PASS, 5 of 5 phases. The runner created snapshot PR #209, read the expected bot replies, discarded the snapshot, and restored the Release Issue to planned.

Changelog input

release-note
None

Additional documentation

This section can be blank.

docs
N/A

[hdamker]

Stability change for regression runner only, no change to RA or Validation. Merging before #345 to see it in action after the merge of #345.