Zipkin

Zipkin is an open source distributed tracing system. It collects timing data from services so that you can find where latency and failures come from in a system made of many services.

Applications are instrumented with a tracer library that reports timing data to Zipkin over HTTP, Kafka or other transports. You can look up a trace by its ID, or search by service, operation name, tags and duration. The interface summarises how much time each service took and whether operations failed. A dependency diagram shows how many traced requests moved between applications, which helps spot error paths and calls to old services.

• Trace search by service, tag and duration
• Summary of time spent in each service
• Dependency diagram between applications
• Reporting over HTTP, Kafka or gRPC
• Storage in memory, Cassandra or Elasticsearch

Backend developers and operations engineers who run service-oriented systems and want a trace store they can host themselves. It is also a good starting point for learning how distributed tracing works.

Zipkin is licensed under Apache-2.0 and the source is on GitHub, so there is no licence fee. Data kept only in memory is lost when the server restarts, so a persistent backend is needed for real retention. Applications must be instrumented before they report anything, and that work is different for each language.