Zipkin

Distributed tracing system for collecting and analyzing timing data across microservices. Helps pinpoint latency issues by gathering span data from instrumented applications and visualizing request flow end-to-end. Supports trace lookup by ID, query by service name, operation, tags, or duration, plus a dependency diagram showing how requests move through the system.

Maintained by OpenZipkin with official instrumenters for Java, Go, JavaScript, Ruby, C#, PHP, and Python. Integrates with Spring Cloud Sleuth, Micrometer Tracing, and OpenTelemetry for drop-in instrumentation. Jaeger and Apache SkyWalking can ingest Zipkin format, so existing clients work with multiple backends.

Key features:

• Span collection over HTTP, Kafka, RabbitMQ, or gRPC transports
• Storage backends: in-memory, MySQL, Cassandra, or Elasticsearch
• B3 propagation for request context across services
• Instrumentation libraries for major languages and frameworks
• Dependency visualization and percentage-of-time breakdowns in the UI

Engineers rely on Zipkin to debug slow requests, trace errors across service boundaries, and understand service dependencies. Common setups include instrumenting API gateways, backend services, and message consumers to trace requests from entry point to database or downstream calls.