Zipkin: Making distributed systems traceable

CurioSwitch engineers contribute to Zipkin and other projects in the OpenZipkin ecosystem.

When a system is made of many services, finding the source of latency or an error becomes difficult. Zipkin gathers timing and metadata recorded by each service and reconstructs the path of a request as a single distributed trace.

Following the journey of an invisible request

Instrumentation libraries record the start, end, and relationships of operations as spans. Reporters send those spans to a collector, storage keeps them available, and the query API and web UI make them searchable and understandable.

Jump directly to a complete request using its trace ID
Search by service, operation, tag, and duration
See latency and failed operations across a trace
Explore service relationships through a dependency diagram

Connecting to different production environments

Zipkin can collect traces through HTTP, Kafka, and other transports. Its storage layer can use in-memory storage, Cassandra, Elasticsearch, and additional backends. This flexibility lets teams introduce tracing in a way that fits the systems they already operate.

CurioSwitch works on technology that turns the inner behavior of complex systems into something teams can understand, troubleshoot, and continuously improve.