An intermittent bug appears only in production on certain iOS Safari versions and you cannot reproduce it locally. Describe your systematic approach to investigate, gather data, and fix this issue given limited access to the production environment.

I’d treat this like a detective problem. First I’d gather telemetry: enable client-side logging and increase sampling in Sentry/Datadog to capture userAgent, OS version, stack, and network timings—initial triage showed the issue impacted ~2% of sessions and 8% of users on iOS Safari 13–14. I’d try remote reproduction with BrowserStack and an actual iPhone using Safari Web Inspector; in my last role I reproduced a rendering race there within 6 hours. If reproduction still fails, I’d add a targeted feature flag (canary at 5%) and a lightweight workaround (debounce, CSS fix) while keeping the flag reversible. I’d coordinate with a small team (3 engineers) to deploy a fix, monitor errors drop by >95% over 24–48 hours, then roll out fully.