About Observability: Logs, Metrics, Traces

Three pillars, RED/USE metrics, OpenTelemetry, tooling. Part of the DevTools Surf developer suite. Browse more tools in the Info / Guides collection.

Use Cases

• SREs building dashboards with RED metrics for service health
• Backend developers adding distributed tracing to microservice architectures
• DevOps teams choosing between Datadog, Grafana, and OpenTelemetry stacks
• On-call engineers using correlated logs and traces to debug production incidents

Tips

• Start with the three pillars overview: logs, metrics, and traces
• Check the RED and USE method sections for choosing the right metrics
• Review the OpenTelemetry setup guide for vendor-neutral instrumentation

Fun Facts

• The term 'observability' comes from control theory, coined by Rudolf E. Kalman in 1960. It entered software engineering vocabulary around 2017 via Charity Majors at Honeycomb.
• OpenTelemetry, formed in 2019 by merging OpenCensus (Google) and OpenTracing (CNCF), is now the second most active CNCF project after Kubernetes.
• Google's Dapper paper (2010) introduced distributed tracing to the industry and directly inspired open-source tools like Zipkin (2012) and Jaeger (2017).

FAQ

Logs, metrics, traces — which first?

Logs are easiest to add, most familiar. Metrics come next for system health (dashboards). Traces last but most valuable — see requests flow across services.

What are RED and USE metrics?

RED for services (Rate, Errors, Duration). USE for resources (Utilization, Saturation, Errors). Both give consistent monitoring frameworks for different layers.

OpenTelemetry?

Vendor-neutral SDK for collecting traces, metrics, and logs. Instrument once, ship to any backend (Jaeger, Datadog, Honeycomb). Industry-standard.

What tools should I pick?

Start with your cloud provider's free tier (CloudWatch, Cloud Monitoring, Azure Monitor). Move to Datadog/Honeycomb when you need more features or the cloud tools become expensive.