# Signals: Traces, Metrics and Logs — OpenTelemetry

Source: https://www.skillbyai.com/en/opentelemetry/f-signals

> Plus shared context.

## Three signals, connected

**Traces** record the path of a request through services as a tree of spans with timings. **Metrics** are aggregated numbers over time (request counts, latency histograms, memory). **Logs** are timestamped records; OpenTelemetry can attach trace and span IDs to them so you can jump from a log line to its trace. All signals share a **resource** describing where they came from (service name, version, host, Kubernetes pod). The specification also covers profiling as a newer signal.

## How the signals connect during an investigation

A typical workflow.

```text
metrics   : error ratio for checkout jumps to 4%      -> alert fires
  |  exemplar / time range
traces    : slow, failing checkout traces show a 3 s span in payment-service -> card-gateway
  |  trace_id
logs      : payment-service log lines with that trace_id show "gateway timeout after 3000 ms"
resource  : all three carry service.name=payment-service, service.version=2.8.1, k8s.pod.name=...
```

## Correlation is the main payoff

Shared IDs and resource attributes turn three separate tools into one investigation path.

**Quiz:** Which signal shows how long each step of a single request took across services?

- [x] Traces
- [ ] Metrics
- [ ] Resource attributes
- [ ] Configuration files

*Answer:* Traces. Spans record timings per step.