Making signals useful
Correlating traces, metrics, and logs; choosing auto, manual, or hybrid instrumentation; building dashboards; and putting the Collector and sampling in the path.
Correlation
Linking traces, logs, metrics, and profiles in Grafana through a shared trace_id and the datasource provisioning that wires it together.
Auto vs. manual vs. hybrid
A decision framework for what auto-instrumentation already covers, what to add by hand, and how the hybrid actually looks in each of the three stacks.
Dashboards
Provisioned dashboards as code, the RED and USE methods, and what a service dashboard needs to answer the first on-call question.
The Collector
Collector pipelines as receivers, processors, and exporters, and why cross-cutting decisions like batching, redaction, and routing belong there instead of in application code.
Sampling
Head sampling versus tail sampling, the tail-sampling Collector config, and what each approach costs you in traces you never see.