Jaeger

Follow one trace across service boundaries

The Collector receives spans emitted by application OpenTelemetry SDKs and relates them into traces. Through the Query Service and web UI, operators can inspect timing, parent-child relationships, and error locations across multiple services.

Sources: [1]

Separate collection, querying, and storage for the deployment

Jaeger v2 uses OpenTelemetry Collector components and lets operators arrange the Collector, Query Service, UI, and storage for their environment. Storage plugins connect the persistence layer, and the README documents support policies for Elasticsearch, OpenSearch, Cassandra, and ClickHouse.

Sources: [1]

Fits investigation of latency and dependencies in service architectures

Jaeger fits teams that need to identify slow segments or downstream failures when one request crosses multiple services. It is also a candidate when OpenTelemetry is the common instrumentation standard from development troubleshooting through production observability.

Separate all-in-one evaluation from production and track compatibility

The Docker all-in-one example combines the UI, Collector, Query, and in-memory storage for quick evaluation. Production deployments must separately design persistent storage, retention, ingestion capacity, redundancy, authentication, and network exposure.

Because Jaeger v2 uses OpenTelemetry Collector components, dependency changes can deprecate configuration options. The project promises at least three months or two minor releases before eligible removal, but upgrades still require checking release notes and configuration compatibility.

Sources: [1][2]