Jaeger

Jaeger is the CNCF's distributed tracing platform — it follows requests across microservices, visualizes the call chain, and pinpoints exactly where latency spikes or errors happen. Originally built at Uber for their massive microservices architecture, it's now the go-to open-source tracing tool.

If you're running microservices on Kubernetes and need to debug "why is this API call slow," Jaeger with OpenTelemetry instrumentation is the standard answer. Zipkin is the simpler, lighter alternative for smaller deployments. Grafana Tempo is the newer option that integrates natively with the Grafana stack. Datadog APM is the commercial "just works" option.

The catch: Jaeger does tracing only — no metrics, no logs. You need Prometheus for metrics and Loki/Elasticsearch for logs, then Grafana to tie it all together. That's a lot of infrastructure. Jaeger v2 embraces OpenTelemetry but the migration from v1 is non-trivial. And for most indie projects with fewer than 5 services, tracing is overkill — structured logging gets you 80% of the debugging value.