A production .NET API needs an urgent hotfix. How would you handle the code and deployment flow?

In production I separate mitigation, repair, and follow-up. Mitigation reduces harm immediately; the hotfix restores correct behavior; later work addresses root cause and wider design without inflating the urgent diff. I canary or stage the artifact where the platform allows it, run a safe synthetic check, and watch the exact failure signal plus overall error rate, latency, saturation, and business outcomes. Rollback criteria are agreed before traffic moves, but rollback is not assumed safe when the hotfix changes data or external effects, so a forward correction or feature disablement may be the recovery path. I preserve logs and timelines, communicate impact and decision points, and avoid manual production edits that cannot be reproduced. After stabilization, the fix is merged forward to every active release line, tags and release notes are reconciled, temporary flags or config are removed, and the regression test remains. The review asks why earlier tests, observability, or release gates missed the defect without turning the incident into blame. That keeps urgency from creating a second, less visible divergence.