Skip to content

RED Metrics

RED (Rate, Errors, Duration) is a service-level monitoring methodology focused on request-driven services.

What are RED Metrics?

RED is a monitoring framework introduced by Tom Wilkie at Grafana Labs, designed specifically for monitoring request-driven services (APIs, microservices, web applications). The three pillars are:

  • 📈 Rate: The number of requests per second your service is handling
  • ❌ Errors: The number/percentage of those requests that are failing
  • ⏱️ Duration: The amount of time those requests take (latency percentiles)

SLOs in this Example

SLO Type Description
ExampleRateSLO Rate Tracks requests per second
ExampleErrorRateSLO Errors Success ratio (2xx, 3xx vs 4xx, 5xx)
ExampleDurationSLO Duration P95 latency
ExampleDurationP99SLO Duration P99 latency (tail)
ExampleAvailabilitySLO Derived Availability from Rate + Errors

Usage

import redmetrics "github.com/grokify/slogo/examples/red-metrics"

// Get individual SLOs
rateSLO := redmetrics.ExampleRateSLO()
errorSLO := redmetrics.ExampleErrorRateSLO()
durationSLO := redmetrics.ExampleDurationSLO()

// Get all SLOs
slos := redmetrics.SLOs()

When to Use RED Metrics

RED metrics are ideal for:

  • Request-driven services: REST APIs, GraphQL APIs, microservices
  • Web applications: Backend services handling HTTP requests
  • Service-to-service communication: Internal API calls
  • User-facing endpoints: Any service where users make requests

Prometheus Queries

Rate

rate(http_requests_total[5m])

Errors

sum(rate(http_requests_total{status=~"2..|3.."}[5m])) / sum(rate(http_requests_total[5m]))

Duration (P95)

histogram_quantile(0.95, rate(http_request_duration_seconds_bucket[5m]))

Duration (P99)

histogram_quantile(0.99, rate(http_request_duration_seconds_bucket[5m]))

Ontology Labels

All RED metric SLOs use these labels:

ontology.LabelFramework:  ontology.FrameworkRED,
ontology.LabelLayer:      ontology.LayerService,
ontology.LabelScope:      ontology.ScopeCustomerFacing,
ontology.LabelAudience:   ontology.AudienceSRE,

Complementary Frameworks

Framework Focus Use Case
USE Metrics Infrastructure CPU, memory, disk, network
Four Golden Signals Hybrid RED + Saturation
DORA Metrics Deployment DevOps performance

References