An Interview With Nikolay Sivko

What does it take to build an observability platform that surfaces root cause insights without overwhelming engineering teams with data? In this episode, Nikolay Sivko, Founder and CEO of Coroot, shares the thinking behind Coroot and how it approaches some of the hardest problems in modern observability.

The conversation covers what sparked the need for Coroot, how it fits into the broader observability market, and where it stands out against existing solutions. Nikolay discusses how metrics, logs, traces, and continuous profiling ship out of the box, and where profiling has made the biggest practical difference for users.

We dig into the technical realities of zero-code instrumentation across every language on Kubernetes, including the trickiest edge cases across runtimes like Java, Go, and Rust. Nikolay also addresses the challenge of collecting comprehensive telemetry without impacting the workloads being observed, and how Coroot manages CPU and memory overhead from eBPF-based instrumentation.

The discussion also covers how Coroot uses AI to surface actionable insights, how it avoids false positives, and how raw telemetry actually gets mapped to a probable root cause under the hood. Nikolay shares the thinking behind the idea that 80% of issues follow common patterns, and how linking logs with other system data helps uncover the remaining 20%.

The episode closes with a look at the future of observability and what Nikolay has learned about building and sustaining an engaged open-source community that has grown to over 6,000 GitHub stars.