Crash Reporting 101

I will discuss how we developed a crash reporting service for mobile applications based on our observability platform, Sage. I will explain the need for a separate service and why the solution that worked well for server applications was insufficient. We will explore which open-source components can be used to implement a similar service and how, using our experience, to make such a service stable and trouble-free.

I will cover why we chose the Sentry SDK, how we learned to group errors, deobfuscate stack traces, and calculate accurate statistics. I will also discuss why processing iOS is more complex and expensive, and what additional challenges we had to address for this platform. In conclusion, I will explain why server applications unexpectedly started connecting to us, where we decided to develop the service further, and our perspective on OpenTelemetry.

This talk will be useful for developers focused on the reliability of mobile applications, providing an "under the hood" look at crash reporting services. It will also benefit engineers responsible for monitoring systems in companies developing mobile applications.