README
¶
RED Metrics Examples
This directory contains OpenSLO examples for RED (Rate, Errors, Duration) metrics, 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 (typically measured as latency percentiles)
Files in this Directory
rate.go
Measures the request rate (throughput) for an API service:
ExampleRateSLO- Tracks requests per second to ensure the service handles expected traffic volume
error.go
Monitors the error rate of requests:
ExampleErrorRateSLO- Tracks the ratio of successful responses (2xx, 3xx) vs errors (4xx, 5xx)
duration.go
Tracks request latency and response times:
ExampleDurationSLO- Measures P95 latency to ensure fast response timesExampleDurationP99SLO- Measures P99 latency for tail latency monitoring
red_test.go
Validation tests for all RED metric 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
Example Prometheus Queries
The examples use Prometheus as the metric source with queries like:
Rate:
rate(http_requests_total[5m])
Errors:
sum(rate(http_requests_total{status=~"2..|3.."}[5m])) / sum(rate(http_requests_total[5m]))
Duration:
histogram_quantile(0.95, rate(http_request_duration_seconds_bucket[5m]))
Complementary Frameworks
- USE Metrics: For infrastructure/resource monitoring (see
../use-metrics/) - Four Golden Signals: Google's framework (similar to RED + Saturation)
- DORA Metrics: For measuring DevOps/deployment performance
References
- The RED Method - Tom Wilkie
- Monitoring Distributed Systems - Google SRE Book
- OpenSLO Specification: https://github.com/OpenSLO/OpenSLO
Documentation
¶
Index ¶
- func AlertConditions() []v1.AlertCondition
- func AlertNotificationTargets() []v1.AlertNotificationTarget
- func AlertPolicies() []v1.AlertPolicy
- func AllObjects() slogo.Objects
- func ExampleAvailabilitySLO() v1.SLO
- func ExampleDurationP99SLO() v1.SLO
- func ExampleDurationSLO() v1.SLO
- func ExampleErrorRateSLO() v1.SLO
- func ExampleRateSLO() v1.SLO
- func SLOs() []v1.SLO
- func SLOsBySetSlug() map[string][]v1.SLO
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func AlertConditions ¶ added in v0.3.0
func AlertConditions() []v1.AlertCondition
AlertConditions returns standard burn rate alert conditions for SLOs. These implement the Google SRE multi-window, multi-burn-rate alerting approach.
func AlertNotificationTargets ¶ added in v0.3.0
func AlertNotificationTargets() []v1.AlertNotificationTarget
AlertNotificationTargets returns example notification targets.
func AlertPolicies ¶ added in v0.3.0
func AlertPolicies() []v1.AlertPolicy
AlertPolicies returns alert policies for the RED metrics SLOs.
func AllObjects ¶ added in v0.3.0
AllObjects returns all OpenSLO objects including SLOs, alert conditions, notification targets, and alert policies.
func ExampleAvailabilitySLO ¶
ExampleAvailabilitySLO is a derived SLO that measures service availability. Availability is calculated from Rate (R) and Error (E) metrics in RED - combining whether the service is handling requests AND those requests are successful.
func ExampleDurationP99SLO ¶
ExampleDurationP99SLO is a SLO that measures P99 latency for more stringent requirements.
func ExampleDurationSLO ¶
ExampleDurationSLO is a SLO that measures request duration (latency). Duration is the "D" in RED metrics - tracking how long requests take to complete.
func ExampleErrorRateSLO ¶
ExampleErrorRateSLO is a SLO that measures the error rate of requests. Error rate is the "E" in RED metrics - tracking the proportion of failed requests.
func ExampleRateSLO ¶
ExampleRateSLO is a SLO that measures the request rate (requests per second) for an API service. Rate is the "R" in RED metrics - tracking how many requests the service handles.
func SLOsBySetSlug ¶ added in v0.3.0
Types ¶
This section is empty.
Source Files