Distributed trace collection and production investigation
Module 1
Jaeger Architecture & Trace Collection
- Jaeger architecture: collector, query service, UI, storage backends
- Jaeger storage: in-memory (dev), Elasticsearch, Cassandra, Badger
- Deployment: Jaeger Operator for Kubernetes, all-in-one binary for development
- OpenTelemetry integration: OTLP receiver in Jaeger — collector pipeline
- Jaeger agent (deprecated) vs direct OTLP export
- Sampling strategies: constant, probabilistic, rate-limiting, remote sampling
- Adaptive sampling: Jaeger's dynamic sampling based on traffic
- Service Performance Monitoring (SPM): RED metrics derived from spans
Module 2
Trace Analysis & Production Workflows
- Jaeger UI: trace search by service, operation, tags, duration
- Trace waterfall view: span hierarchy, timing, gap analysis
- Critical path analysis: identifying the slowest path through a trace
- Comparing traces: spotting regressions between deployments
- Tagging best practices: span attributes for business context
- Dependency graph: service topology from trace data
- Integrating Jaeger with Grafana: Grafana Tempo as Jaeger-compatible backend
- Production capacity: storage sizing for retention periods
- Multi-tenancy: Jaeger with tenant-scoped storage
What your team walks away with
Teams who can deploy Jaeger, instrument services with OpenTelemetry, and use trace analysis to diagnose production problems that logs alone cannot reveal.
- Deploy Jaeger in Kubernetes using the Operator and configure storage for production retention
- Collect traces via OTLP from OpenTelemetry-instrumented services
- Configure adaptive sampling to balance coverage and storage cost
- Use the Jaeger UI to identify slow spans, gaps, and error paths in distributed requests
Book the Jaeger training
Works best as part of the OpenTelemetry & Distributed Tracing training. Can be delivered standalone for teams who already instrument with OTel.
Get in touch