Phase 2: Tools & Resources
Tools for measuring engineering impact and building ROI narratives.
Apache DevLake (with gh-devlake helper)
| Setup | High — Docker or Azure, multi-tool config |
| Best for | Correlating Copilot adoption with DORA delivery metrics |
| Data | Full history in Apache DevLake's backing database |
| Cost | Free locally (Docker). ~$30-50/mo on Azure |
Apache DevLake is the platform that ingests GitHub, Copilot, and delivery data into a common model and exposes pre-built Grafana dashboards. gh-devlake is the CLI helper used to deploy and configure that stack.
Includes pre-built Grafana dashboards:
- Adoption Dashboard — DAU/WAU/MAU, acceptance rates, seat effectiveness
- Impact Dashboard — Correlates adoption tiers with PR velocity, deployment frequency, CFR, MTTR
gh extension install DevExpGBB/gh-devlake
gh devlake deploy local --dir ./devlake
cd devlake && docker compose up -d
gh devlake configure full
# Open Grafana at http://localhost:3002
Existing BI stack (Splunk, Power BI, Tableau, and similar)
| Setup | Medium — requires API ingestion, data modeling, and dashboard design |
| Best for | Teams that already standardize on Splunk, Power BI, Tableau, Fabric, or another analytics platform |
| Data | Depends on your retained API exports, GitHub data, CI/CD data, incidents, and survey responses |
| Cost | Existing platform and licensing costs |
Use the Copilot Usage Metrics API and dashboard exports as the Copilot input, then combine them with GitHub, CI/CD, incident, and survey data in the analytics stack your organization already trusts.
copilot-metrics-tools (Premium Request Usage)
| Setup | Low — bash script or Next.js app |
| Best for | Understanding premium request costs and top consumers |
| Data | Point-in-time reports |
| Cost | Free |
Analyzes premium request consumption by user and model. Tracks included vs billed requests.
export GITHUB_TOKEN="ghp_..."
export GITHUB_ENTERPRISE="your-enterprise"
./scripts/copilot-report.sh --year 2026 --month 2 > report.json
cp .env.local.example .env.local # edit with your token
npm install && npm run dev
# Open http://localhost:3000
Comparison
| Tool | Focus | Setup | DORA | Cost Analysis |
|---|---|---|---|---|
| Apache DevLake | Adoption → Impact correlation | High | Yes | No |
| Existing BI stack | Custom dashboards and correlation | Medium | Yes | Possible |
| copilot-metrics-tools | Premium request costs | Low | No | Yes |
→ Full catalog including Phase 1 tools: Tool Catalog
What to do next:
- Use Apache DevLake for a prebuilt DORA correlation stack
- Already use Splunk or Power BI? Start with your existing analytics stack and the Analytics-Ready Playbook
- Run copilot-metrics-tools to check premium request spend