Tracking Engineering Velocity
Use OpsBrief data to understand your engineering team's release cadence, incident frequency, and overall velocity.
Why Track Engineering Velocity?
Engineering velocity isn't just about shipping fast - it's about understanding your team's rhythm and identifying opportunities to improve.
- Predictability: Know when features will ship
- Resource planning: Staff appropriately
- Process improvement: Identify bottlenecks
- Team health: Spot burnout indicators
Metrics You Can Track
Release Frequency
How often you ship to production
How to measure: Count 'Release' events per week/month from your dashboard
Benchmark: High performers: 1+ per day
Incident Rate
How often things break
How to measure: Count 'Incident' events per week/month
Benchmark: Track trend over time, aim to reduce
Deployment Frequency
How often you deploy (including staging)
How to measure: Count 'Deployment' events per week
Benchmark: DORA elite: multiple per day
Change Fail Rate
Percentage of releases causing incidents
How to measure: Incidents / Releases in same period
Benchmark: DORA elite: 0-15%
Using the Dashboard for Analysis
Weekly Review
Every Monday, open the dashboard and filter by 'Last 7 days'. Count releases and incidents.
Dashboard → Filter: Last 7 days → Note counts
Monthly Trends
Use the calendar view to see patterns. Are releases bunching up before deadlines?
Dashboard → Calendar View → Look for patterns
Export for Analysis
Download events as CSV for spreadsheet analysis or reporting.
Dashboard → Export → CSV
Interpreting the Data
Pattern: Releases decreasing
Possible causes:
- •Growing tech debt
- •Team burnout
- •Process bottlenecks
Actions:
- →Invest in developer tooling
- →Schedule tech debt sprints
Pattern: Incidents increasing
Possible causes:
- •Moving too fast
- •Missing tests
- •Infrastructure issues
Actions:
- →Add CI/CD checks
- →Review deployment process
Pattern: Friday deployments failing
Possible causes:
- •Rushed releases
- •Reduced team availability
Actions:
- →Freeze Fridays
- →Require Friday PR reviews
Best Practices
Don't gamify the metrics
Velocity metrics should inform decisions, not become targets. Goodhart's Law applies.
Context matters
A week with zero releases during a refactor sprint is fine. Track trends, not single data points.
Combine with qualitative data
Use OpsBrief data alongside team feedback, sprint retros, and 1:1s.
Share with the team
Transparency builds trust. Share velocity reports in standups.
Sample Weekly Report
# Engineering Velocity Report
## Week of Dec 2-8, 2024
### Summary
- Releases: 8 (↑ from 6 last week)
- Incidents: 2 (↓ from 4 last week)
- Deployments: 23
- Change Fail Rate: 8.7%
### Highlights
- v2.3.0 shipped on Tuesday
- Database migration completed
### Incidents
- API latency spike (resolved in 45min)
- Auth service timeout (resolved in 2hr)