pixie

Why we included this project

Pixie collects telemetry from Kubernetes clusters using eBPF, so teams can see what their pods are actually doing without adding instrumentation or changing code. It captures full-body application requests, service maps, resource and network metrics, and profiling data, all viewable from a web UI or through scripts. That makes it handy for debugging a slow endpoint, triaging a noisy node, or tracing cross-service traffic, since the drill-down views cover pod state, flame graphs, and raw request bodies. Collection happens automatically and data stays inside the cluster, so it fits into daily work instead of demanding upfront setup. Teams that want observability without rebuilding or redeploying their applications will get the most out of it.