Services (RED & SLO)

The Services page shows the golden signals — Rate, Errors, Duration — plus an SLO error budget and multi-window burn rate, for any service. It's a Grafana Scenes page whose queries are generated from your cluster's traffic profile, so it adapts to whatever ingress/mesh you run.

What you get

• Request rate, Error %, Availability (SLI) as stats.
• Latency p50 / p90 / p99 (percentiles where the stack exposes a histogram; an average fallback for stacks like HAProxy that don't).
• Error budget remaining (30d) against a configurable SLO target.
• Burn rate (1h / 6h) — fast-burn when both windows exceed 14.4×, the Google-SRE multi-window signal.

Pick your stack

Set the Traffic / RED source in the cluster config — or click Auto-detect to probe your Prometheus. Supported out of the box: Istio, Linkerd, Cilium/Hubble, ingress-nginx, Traefik, HAProxy, Envoy (Gateway API / Contour / Emissary), Kong, Consul, Kuma, plus generic Prometheus-client HTTP and OpenTelemetry HTTP — or a fully custom metric/label mapping.

See Traffic stack profiles for the exact metrics each profile uses and how to map a custom stack.