Learn how Zipkin, the open-source distributed tracing system from Twitter engineering, tracks latency across microservices. Features, architecture, and FAQs.
Zipkin is an open-source distributed tracing system that helps engineering teams troubleshoot latency problems in microservice architectures by collecting timing data as requests move across services.
The project originated at Twitter in 2012, where engineers built it to diagnose performance issues in the company's rapidly scaling service infrastructure. In mid-2015, the project moved from Twitter to OpenZipkin, an independent community organization that maintains it on GitHub today.
Zipkin is released under the Apache License 2.0, so it can be self-hosted, modified, and used commercially free of charge.
Applications send trace data to Zipkin after being instrumented with a compatible tracer or client library. Zipkin supports multiple transport options for receiving this data, including HTTP, Kafka, Apache ActiveMQ, gRPC, and RabbitMQ, and can store traces in memory or in production backends such as Apache Cassandra or Elasticsearch.
The Zipkin UI allows engineers to search traces by trace ID or by filtering on service name, operation name, tags, and duration, making it possible to quickly isolate a slow or failing request.
Zipkin automatically summarizes key performance data, such as the percentage of time spent per service and operation failure rates, and generates dependency diagrams that visualize how requests flow through an application, helping teams identify bottlenecks and deprecated service paths.
Zipkin is completely free and open source under the Apache License 2.0. There are no paid tiers, subscriptions, or official commercial editions.
Costs are limited to the infrastructure a team chooses for hosting Zipkin's collector and storage backend, such as servers running Cassandra or Elasticsearch.
Support is community-driven through GitHub issues and community chat channels rather than a paid support contract.
Zipkin is used to trace requests as they move across microservices, helping engineers diagnose latency and failure points in distributed systems.
Yes. Zipkin is open source and released under the Apache License 2.0, with no cost to download, self-host, or modify.
Zipkin was created by Twitter engineers in 2012 and has been maintained by the independent OpenZipkin community since 2015.
Applications are instrumented with a tracer library, then send trace data to Zipkin over HTTP, Kafka, Apache ActiveMQ, gRPC, or RabbitMQ.
Zipkin can store traces in memory for testing, or in production backends such as Apache Cassandra or Elasticsearch.
Both are open-source distributed tracing systems; Zipkin predates Jaeger and originated at Twitter, while Jaeger was created at Uber, with each having its own storage and UI implementation.
Zipkin itself is a self-hosted open-source project without an official managed cloud offering.
Yes, many OpenTelemetry-compatible instrumentation libraries can export trace data to a Zipkin backend.