observability

You can set up OpenTelemetry tracing and Prometheus metrics by applying auto-instrumentation for minimal code changes and configuring OTLP exporting to collectors. This provides end-to-end visibility by collecting and correlating logs, metrics, and traces across your distributed services.

Correlating logs and traces across service boundaries requires structured logging schemas and OpenTelemetry manual spans. By emitting JSON logs and OTLP traces with shared context, you achieve end-to-end visibility and link business logic directly to system behavior.

Yes, this approach supports defining SLOs, alerting rules, and Grafana dashboards using Prometheus metrics. It enforces metric naming conventions with low cardinality to ensure your dashboards remain performant and reliable as your architecture scales.

Yes, you can use auto-instrumentation to emit logs, metrics, and traces with minimal code changes. This allows you to start collecting observable data immediately by enabling JSON logs, Prometheus metrics, and OpenTelemetry tracing across your services.

High cardinality causes issues with Prometheus metrics by exploding the amount of stored time series data, which degrades query performance. You avoid it by enforcing metric naming conventions with low cardinality when designing your alerting rules and Grafana dashboards.