Observability
Tracing Journeys Without Noise
Keep traces useful under load with sampling strategies that preserve causality.
4 weeks · 18h · Self-paced + mentor checkpoints · BRL 1.420 (informational)
Program narrative
From trace context propagation to tail sampling, you build a tracing plan that survives peak traffic and still answers "why slow" during incidents.
What ships inside
- Context propagation checklist
- Tail vs head sampling scenarios
- TraceQL-style query patterns
- Cardinality guardrails
- Service map sanity tests
- Trace-driven retro template
- Noise budget worksheet
Outcomes we expect you to demo
- Ship a sampling policy doc tied to SLO tiers
- Cut redundant spans by 30% in a sample workload
- Facilitate a trace-based incident narrative
Mentor on point
Otávio Mendes
Observability mentor with background in high-volume retail peaks.
FAQ — including what we skip
Examples rotate between Tempo-style and vendor-neutral OTLP pipelines; you adapt to your storage.
Experience notes
“Tail sampling labs mirrored our Black Friday rehearsal—finally a course that names cardinality explicitly.”
— Igor