How do you approach debugging a system you did not build?

When debugging an unfamiliar system, I start by understanding the high level architecture and the data flow between components.

Next, I reproduce the issue in a controlled environment and use logs, metrics, and traces to narrow down where the failure occurs. I form hypotheses and validate them incrementally rather than making broad changes.

Throughout the process, I document findings and communicate with team members who have context. This structured approach helps reduce guesswork and speeds up root cause identification.