Troubleshooting Microservice Architectures

What does it actually take to troubleshoot distributed systems made up of hundreds or even thousands of interconnected services? At Devoxx Poland, Peter Zaitsev, Coroot Co-Founder, tackled the core observability challenges facing SRE and DevOps teams working at scale.

Peter addressed the foundational questions that matter most: what information do you actually need about each service, and how do you design dashboards that give engineers an immediate, clear picture of system health rather than raw data they have to interpret under pressure.

The session explored how observability signals, including metrics, logs, traces, and profiles, each serve a distinct role and what unique insights each one provides. Rather than treating these signals as interchangeable, Peter examined when to reach for each one and how to use them together effectively.

Peter also covered practical approaches to automating root cause analysis, weighing the trade-offs between fully automated and partially automated solutions, and where human judgment remains essential in the troubleshooting process.