dora-measurement

Skill: DORA Measurement

Load trigger: "load dora-measurement skill" > DORA: AI Capability 2 (Healthy data ecosystems) + Capability 7 (Quality internal platforms) Token cost: Medium (queries external endpoints)

Purpose

Translate uFawkesObs telemetry into the four DORA delivery metrics and a plain-language ROI signal. Bridges the gap between "substrate is running" and "we have DORA numbers."

Dependency: uFawkesObs must be running and ingesting events. If deployment event sources are not yet wired from uFawkesPipe, proxy metrics are used and flagged explicitly. Never produce metrics without flagging proxy usage — that would be misleading data.

Pre-conditions

# Verify uFawkesObs is running
curl -s "${PROMETHEUS_URL}/-/healthy" | grep -q "Prometheus" || echo "ERROR: Prometheus not healthy"
curl -s "${LOKI_URL}/ready" | grep -q "ready" || echo "ERROR: Loki not ready"
curl -s "${GRAFANA_URL}/api/health" | jq '.database' | grep -q "ok" || echo "ERROR: Grafana not healthy"

# Required environment variables
: "${PROMETHEUS_URL:?Set PROMETHEUS_URL (e.g. http://localhost:9090)}"
: "${LOKI_URL:?Set LOKI_URL (e.g. http://localhost:3100)}"
: "${GRAFANA_URL:?Set GRAFANA_URL (e.g. http://localhost:3000)}"
: "${MEASUREMENT_WINDOW_DAYS:=30}"
: "${REPO:?Set REPO (e.g. paruff/uFawkesObs)}"

The Four DORA Delivery Metrics

1. Deployment Frequency

How often code is successfully deployed to production.

Primary query (Prometheus):

# Deployments per week over the measurement window
rate(deployment_events_total{repo=~"REPO", status="success"}[${WINDOW}d]) * 604800

Proxy metric (if deployment events not yet wired — flag proxy_metrics: true):

# PR merge rate as proxy for deployment frequency
rate(github_pr_merged_total{repo=~"REPO"}[${WINDOW}d]) * 604800

DORA tier thresholds:

Tier	Value
Elite	On-demand (multiple deploys/day)
High	1/week to 1/day
Medium	1/month to 1/week
Low	Less than 1/month