Jaeger

Jaeger is an open-source tracing project for teams diagnosing requests that cross distributed services. Its v2 release is the sensible starting point: v1 reached end of life on December 31, 2025.

Overview

Jaeger helps locate performance bottlenecks, trace root causes, and examine service dependencies. It covers distributed and transaction tracing, plus dependency mapping, without a software charge. The trade-off is operational ownership: deployment is self-hosted, and storage capacity and scale depend on the infrastructure you choose.

Key features

Jaeger accepts trace data through OpenTelemetry Protocol (OTLP) and Zipkin spans over HTTP in Thrift, JSON v1/v2, and Protobuf formats. That breadth suits teams already emitting either protocol, rather than requiring a single ingestion path.

The UI includes System Architecture and Deep Dependency Graph service views. Service Performance Monitoring derives aggregate metrics from traces and displays them as time-series charts, helping operators move from individual requests to service-level patterns.

Sampling supports head-based configuration managed centrally, with static or adaptive settings, as well as tail-based sampling. This gives teams options for controlling trace collection, but makes sampling policy part of the operational work.

Primary distributed storage integrations are Cassandra, Elasticsearch, and OpenSearch. ClickHouse is experimental and gated, so it is not the equivalent of a primary supported backend. Jaeger documents TLS and mutual TLS for several component connections, and bearer-token authentication between the browser and UI.

Jaeger v2 ships as executable binaries and container images. Its all-in-one setup combines collector and query components in one process, but uses transient in-memory storage; it is suited to a compact setup, not durable retention. Binary releases include SHA-256 checksum files and cryptographic signatures.

Pricing

Jaeger — free. The plan is 100% open source and self-hosted, with no free-trial distinction. There are no published capacity or scale allowances: both depend on the infrastructure and storage you provide. This is a strong fit for teams able to provision and operate that stack; it is not a hosted service with included managed capacity.

Platforms

Jaeger supports API and web access, self-hosted deployment, and Linux, macOS, and Windows. Downloadable binaries are available for all three desktop/server operating systems, alongside container images.

Who it's for

Jaeger suits engineering teams that need distributed traces, service dependency views, and trace-derived performance metrics, and can take responsibility for storage and deployment. It is less suitable for teams seeking real user monitoring or a managed observability service: real user monitoring is not included, and self-hosting is the deployment model.

Pros and cons

• Protocol flexibility: OTLP and multiple Zipkin span formats support varied existing instrumentation.
• Useful investigation views: Service graphs and trace-derived time-series metrics connect request-level diagnosis to broader service patterns.
• Operator control: Multiple sampling approaches and primary storage choices leave configuration in the team's hands.
• Infrastructure burden: Teams provide storage and scale themselves; the all-in-one option's in-memory storage is transient.
• Version transition matters: v1 no longer receives updates, so new deployments should use v2.

Alternatives

Application Performance Monitoring and Distributed Tracing Tools are useful category lists for comparing the field.

• Middleware is worth considering for teams that want a freemium option with a free trial and published free-tier allowances for data, RUM sessions, synthetic checks, and browser runs.
• Grafana Cloud Application Observability fits readers looking for a freemium hosted alternative, with a free tier billed as always free and a Pro plan priced by host hour plus a monthly platform fee.
• Apache SkyWalking is another free, open-source, self-hosted observability platform for readers comparing self-managed options.
• Honeycomb may suit teams seeking a freemium service with a free trial and published event and metrics allowances on its free plan.
• ManageEngine Applications Manager is an option for readers wanting application and server monitoring with a free plan capped at five apps or servers.
• New Relic Kubernetes Monitoring is a freemium alternative for readers prioritizing Kubernetes monitoring and a free data-ingest allowance.
• SigNoz offers a self-hosted community plan as well as a cloud option, which may fit teams weighing self-management against hosted deployment.
• Uptrace is a freemium alternative with a cloud free allowance for monthly data, suitable for teams preferring a hosted option over managing Jaeger's storage.

Verdict

Choose Jaeger if your team needs capable distributed tracing and service-level investigation without software fees, and is prepared to own deployment, storage, and scale. Look elsewhere if you need real user monitoring or want a managed service to take that infrastructure work off your hands.