dora-metrics

name	dora-metrics
description	DORA metric queries for uFawkesObs: PromQL expressions for deployment frequency, lead time, change failure rate, and FDRT sourced from deployment spans processed by the OTel Collector. Load when writing DORA dashboards or interpreting DORA metric data in Prometheus.
license	MIT
compatibility	Claude Code, GitHub Copilot, OpenCode, Cursor, Codex, Gemini CLI
metadata	{"author":"paruff","suite":"uFawkesObs"}

Skill: DORA Metrics (uFawkesObs)

How DORA Metrics Reach Prometheus

Services emit OTLP span events → OTel Collector → spanmetrics processor converts spans to Prometheus counters → Prometheus scrapes → Grafana queries.

Required spans from services:

deployment.completed  { service, version, environment, duration_ms, status }
deployment.failed     { service, version, environment, error }
incident.opened       { service, severity }
incident.resolved     { service, severity, duration_ms }

If these spans are not emitted: DORA metrics will be absent from Prometheus. Verify with: curl -s http://localhost:9090/api/v1/label/__name__/values | grep deployment

PromQL Queries

Deployment Frequency (deployments per day)

increase(deployment_completed_total{environment="prod"}[1d])

Change Failure Rate

sum(rate(deployment_failed_total{environment="prod"}[7d]))
/
sum(rate(deployment_completed_total{environment="prod"}[7d])) * 100

Failed Deployment Recovery Time (FDRT)

# Average recovery duration in seconds
avg(incident_duration_seconds{severity=~"critical|high"})

Lead Time (proxy — PR merge to deployment)

# Requires deployment spans to carry commit_timestamp attribute
# If available:
avg(deployment_lead_time_seconds{environment="prod"})

Note: Lead time is the hardest to measure via spans alone. If deployment_lead_time_seconds is absent from Prometheus, lead time tracking requires integration with the GitHub API or DevLake (the approach used in full fawkes deployments).

DORA Thresholds for Alert Rules

Metric	Warning	Critical
Change failure rate	> 15%	> 30%
FDRT	> 1 day	> 7 days
Deployment frequency	Declining trend	Zero for > 7 days

Grafana DORA Dashboard Variables

"templating": {
  "list": [
    {
      "name": "service",
      "type": "query",
      "datasource": { "uid": "prometheus" },
      "query": "label_values(deployment_completed_total, service)"
    },
    {
      "name": "environment",
      "type": "query",
      "datasource": { "uid": "prometheus" },
      "query": "label_values(deployment_completed_total, environment)",
      "current": { "value": "prod" }
    },
    {
      "name": "range",
      "type": "interval",
      "options": ["1d", "7d", "30d", "90d"],
      "current": { "value": "30d" }
    }
  ]
}

OTel Collector spanmetrics Processor Config

To convert deployment spans to Prometheus metrics, the collector needs the spanmetrics connector:

# In config/otel-collector-config.yaml
connectors:
  spanmetrics:
    histogram:
      explicit:
        buckets: [10ms, 100ms, 500ms, 1s, 5s, 30s]
    dimensions:
      - name: service
      - name: environment
      - name: version
      - name: status
    metrics_flush_interval: 15s

service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: [memory_limiter, batch]
      exporters: [otlp/tempo, spanmetrics]
    metrics/from-spans:
      receivers: [spanmetrics]
      exporters: [prometheusremotewrite]

Verify spanmetrics is working:

curl -s http://localhost:9090/api/v1/query?query=calls_total | python3 -m json.tool
# Should return span-derived metrics if spans are being received