Sentry triage fix-effectiveness verification: cross-reference deploy commits with new event streams

After deploying fixes identified in Sentry triage, the next triage run should cross-reference fix commits against new event streams to verify effectiveness. Pattern:

From healthcheck, get the deployed git_commit
git log --oneline <prev_deploy>..<current_deploy> to identify fix commits in the range
For each fix commit, check whether the targeted Sentry issue cluster has new events on the current release
Classify results: resolved (no new events), reduced (fewer events, possibly new fingerprint/shortId), partial (some variants fixed, others not), ineffective (same rate)

Key insight from this run: a fix (60e9693b4) that patched the edit-apply code path for action-subtype validation was partial — the same bug existed in the direct-build code path, which bypasses the edit-apply pipeline. The fix eliminated edit-apply errors but direct-build errors continued. Always audit parallel code paths for the same bug.

Additional signal: when a fix changes error-handling code (e.g. SSR handleFetch hardening), the old Sentry issue shortIds stop receiving events but NEW shortIds appear with different chunk hashes in the minified culprit. The new shortIds may have lower event rates, confirming the fix reduced severity even if it didn't eliminate the underlying network failures.