Jaeger is an open-source distributed tracing tool that helps engineers trace requests as they move across microservices.
Jaeger is an open-source, end-to-end distributed tracing platform that helps engineers monitor and troubleshoot requests as they move through complex microservice systems. Originally built by Uber and now part of the Cloud Native Computing Foundation, Jaeger connects the dots between the many services a single request touches. For anyone running production systems with more than a handful of services, including AI agent pipelines that chain multiple tool calls, Jaeger turns an opaque request into a visual map you can actually debug.
Jaeger tracks a request from the moment it enters your system to the moment it finishes, across every service it touches along the way.
Trace ID: every incoming request gets a unique trace ID that gets passed along to every service it calls, so all the work done for that one request can be tied back together.
Spans: each unit of work within a trace, like a database query or an API call, is recorded as a span. A span has a name, a start time, a duration, and a link to its parent span.
Client instrumentation: language-specific libraries create spans in your application code with minimal manual setup, and Jaeger now recommends using OpenTelemetry SDKs to generate this data.
Collector and storage: the Jaeger collector receives spans from your services, validates and processes them, then writes them to a storage backend like Elasticsearch or Cassandra.
Jaeger UI: the query service and UI let you search traces by service, operation, or duration, then drill into a visual timeline showing exactly where time was spent and where errors occurred.
Microservice performance dashboards: instrument every service in your stack so you can see request latency broken down by service, spotting exactly which one is slowing down the whole chain.
Root cause analysis for production incidents: trace a failed or slow request back through every hop it took, cutting incident response time from hours of log searching to minutes of trace inspection.
AI agent pipeline debugging: trace multi-step agent workflows where one request triggers several tool calls or LLM invocations, so you can see which step introduced latency or failed.
Service dependency mapping: generate a visual graph of how your services call each other, useful for onboarding new engineers or planning safe changes to a shared service.
Local development testing: run Jaeger’s all-in-one Docker image locally to catch tracing issues and latency problems before a service ever reaches production.
Jaeger is used to trace requests as they travel through a distributed system made up of many microservices. Engineers use it to spot performance bottlenecks, understand how services depend on each other, and trace a failure back to the exact service and operation that caused it, which speeds up debugging in complex systems.
Yes. Jaeger is fully open-source and free to run, hosted under the Cloud Native Computing Foundation. You can self-host it using Docker or Kubernetes with no license cost. Some vendors offer managed or enhanced versions of Jaeger-compatible tracing as a paid service, but the core project itself is free.
OpenTelemetry instruments applications and generates trace data, but it does not store or display it. Jaeger is the backend that receives that trace data, stores it, and provides a UI to search and visualize it. In most modern stacks, OpenTelemetry captures the traces and sends them to Jaeger for analysis.
Braintrust is an AI observability and evaluation platform that helps engineering teams trace, test, and monitor LLM applications in production.
Observability
Helicone is an open-source AI gateway that logs, traces, and optimizes every LLM API call with one line of code.
ObservabilityLangSmith is LangChain's framework-agnostic platform for tracing, evaluating, and deploying AI agents and LLM applications in production.
Observability
Prometheus is an open-source monitoring system that scrapes metrics from your services on an interval, stores them as time-series data, and lets you query and alert on them with PromQL.
ObservabilityUpdates from the AI world — what shipped, what we’re using in production, and what’s worth your attention. Two emails a month, no spam.
