sysdesign-monitoring-4-golden-signals

The four golden signals for monitoring are latency, traffic, errors, and saturation. Defining these signals per service surface establishes clear observability boundaries to guide thresholding, alerting, and incident response for distributed architectures and APIs.

Configuring per-surface role assignment for alerts prevents paging on every spike. You assign specific roles like page, ticket, or dashboard to each signal surface, ensuring alerts scale appropriately and only trigger urgent incidents for critical threshold breaches.

Yes, this monitoring approach works with Prometheus, Datadog, CloudWatch, or OpenTelemetry. It requires a metrics pipeline to apply the four golden signals to live services, establishing observability boundaries and guiding dashboards for distributed architectures.

Observability guidance for burn-rate alerts and downstream saturation considerations provides end-to-end visibility. You apply these concepts alongside the four golden signals to track error budget consumption and detect cascading saturation issues across service dependencies.

Defined ownership with runbooks is required to anchor monitoring to the four golden signals effectively. This principled approach ensures teams establish clear observability boundaries, preventing noisy dashboards and missed outages by assigning accountability for incident response.