I have a scenario where a process normally takes 20 minutes. How would you troubleshoot it?

1. Define the problem – Confirm the 20‑minute duration is consistent and identify the exact process or job.
2. Collect metrics – Look at CPU, memory, I/O, network usage, and any application logs around the execution window.
3. Identify bottlenecks – Use profiling tools (e.g., top, perf, database query plans, or application profilers) to see which step consumes the most time.
4. Check external dependencies – Network calls, database locks, or third‑party services can add latency.
5. Review recent changes – Code commits, configuration changes, or infrastructure updates that coincide with the slowdown.
6. Reproduce in a controlled environment – Run the job with a smaller dataset or in a staging environment to isolate the issue.
7. Apply fixes iteratively – Optimize the slowest step (e.g., add indexes, batch operations, or parallelism) and re‑measure.
8. Document findings – Record the root cause and the solution for future reference.

This systematic approach ensures you pinpoint the root cause rather than applying ad‑hoc fixes.