Jaeger

Overview

Jaeger is an open-source platform for tracing requests across complex distributed systems. It suits teams that can operate their own observability infrastructure and need to investigate bottlenecks, root causes, and service dependencies.

Its tracing and service-graph capabilities make it a focused option for diagnosing distributed workflows, but storage, capacity, and scale are yours to manage. New deployments should use Jaeger v2: v1 reached end of life on December 31, 2025, and no longer receives updates.

Key features

Jaeger accepts trace data through OpenTelemetry Protocol (OTLP) and can ingest Zipkin spans over HTTP in Thrift, JSON v1/v2, and Protobuf formats. That gives teams using either protocol a way to feed traces into Jaeger without making its own instrumentation format the only option.

For investigation, the UI provides System Architecture and Deep Dependency Graph views. Service Performance Monitoring also turns trace data into aggregate metrics and time-series charts. These tools help teams move from an individual request to broader service relationships and performance patterns; they do not replace the work of running the storage and collection infrastructure.

Sampling includes head-based options with centralized remote configuration, which can be static or adaptive, as well as tail-based sampling. This gives operators choices about which traces to retain, though deciding how to configure sampling remains an operational responsibility.

Primary distributed storage integrations are Cassandra, Elasticsearch, and OpenSearch. ClickHouse support is experimental and behind a feature gate, so it is not on the same footing as the primary backends. Security documentation covers TLS and mutual TLS for several component connections, plus bearer-token authentication between the browser and UI.

Jaeger v2 is distributed as executable binaries and container images. Its all-in-one setup runs collector and query components together, but uses transient in-memory storage; it is suitable for a lightweight start, not durable trace retention. Binary releases include SHA-256 checksums and cryptographic signatures. The project has been active since 2016 and describes itself as cloud native and fully open source; community help is available through CNCF Slack, GitHub issues and discussions, and Stack Overflow.

Pricing

Jaeger — free. The plan is 100% open source and self-hosted. There is no software price, but storage capacity and scale depend on the infrastructure you choose and operate. Jaeger has no free trial because its plan is free.

This fits teams that want control over deployment and can provide the supporting infrastructure. It is less suitable for readers seeking a managed service or a predictable, bundled storage allowance: the plan leaves those costs and operational decisions to the user.

Platforms

Jaeger supports API, Linux, macOS, self-hosted, web, and Windows. Downloadable desktop binaries are available for Linux, macOS, and Windows, while the deployment model is self-hosted.

Who it's for

Jaeger is a strong fit for engineering teams tracing distributed services that want open-source tooling, protocol flexibility, service graphs, and control over where trace data is stored. It is especially practical when the team already operates compatible storage such as Cassandra, Elasticsearch, or OpenSearch.

Look elsewhere if you need real user monitoring: Jaeger does not provide it. Also consider another option if your priority is managed hosting, built-in storage quotas, or a supported v1 installation; v1 no longer receives updates.

Pros and cons

• Pros: No software charge and a fully open-source, self-hosted deployment let teams control their tracing stack.
• Pros: OTLP and several Zipkin span formats support integration with more than one tracing protocol.
• Pros: Service graphs, trace-derived performance charts, and multiple sampling approaches support both request-level and service-level investigation.
• Cons: Storage capacity and scale rely on infrastructure the user provides and operates.
• Cons: The all-in-one setup stores data only in transient memory, limiting its role to non-durable use.
• Cons: Jaeger v1 is end of life and receives no updates, making v2 the relevant choice for new deployments.
• Cons: ClickHouse is experimental behind a feature gate, unlike the primary supported storage backends.

Alternatives

For another free, self-hosted open-source observability platform, consider Apache SkyWalking. SigNoz offers a self-hosted community plan as well as a cloud option; its community plan puts infrastructure, storage, scaling, upgrades, and backups on the user.

If managed pricing with a stated ingestion allowance matters, Uptrace has a cloud plan with 50 GB free per month, then $0.075/GB, and unlimited users, services, and hosts. OpenZipkin is another open-source project centered on Zipkin code, documentation, specifications, and support.

For a hosted free tier with explicit trace and retention limits, Tracekit offers 200K traces per month, seven-day retention, and one team member on its free plan. Grafana Tempo is another free option.

Honeycomb has a free plan capped at 20M events per month, 100M metrics data points per month, and two triggers. Grafana Cloud Traces offers an always-free plan with 50 GB ingested per month and 14-day retention.

Compare more options in Distributed Tracing Tools and Application Performance Monitoring.

Verdict

Choose Jaeger if your team needs open-source distributed tracing and can run the storage and deployment infrastructure itself. Its protocol support, service views, and sampling options make it a capable choice for troubleshooting complex systems without a software fee. Look elsewhere for managed operations, bundled retention, or real user monitoring.