Skip to content

Module 2: DORA Metrics - The North Star

Belt Level: 🥋 White Belt Duration: 60 minutes Prerequisites: Module 1 completed DORA Capabilities: Monitoring and Observability, Continuous Delivery

1. Learning Objectives (3 minutes)

What You'll Learn

By the end of this module, you will be able to:

  • ✅ Explain the Four Key Metrics and why they predict software delivery performance
  • ✅ Differentiate between Elite, High, Medium, and Low performers using data
  • ✅ Calculate each DORA metric for your team
  • ✅ Interpret DORA metrics dashboards and identify improvement opportunities
  • ✅ Understand how Fawkes automates DORA metrics collection
  • ✅ Articulate the business impact of improving these metrics

Why It Matters

The Research: The DORA (DevOps Research and Assessment) team spent 9 years studying 32,000+ organizations to answer one question:

"What separates high-performing software teams from everyone else?"

The Discovery: Just four metrics predict organizational performance better than any other measures. Organizations that excel at these metrics are:

  • 2x more likely to exceed profitability goals
  • 2x more likely to exceed productivity goals
  • 2x more likely to exceed customer satisfaction goals
  • 50% more likely to have higher market share

Your Opportunity: These aren't vanity metrics—they're predictive indicators of success. Understanding and improving them is literally your competitive advantage.

Success Criteria

You've mastered this module when you can:

  • Explain each metric to a non-technical executive in business terms
  • Look at a DORA dashboard and immediately spot problems
  • Calculate metrics for your own team
  • Recommend specific improvements based on metric trends
  • Understand how platform engineering improves all four metrics

2. Theory & Concepts (15 minutes)

📺 Video: The Four Key Metrics Explained (7 minutes)

[VIDEO PLACEHOLDER] > See detailed script in supporting document

The Four Key Metrics

DORA identified four metrics that matter most for software delivery performance:

1. 🚀 Deployment Frequency (DF)

Definition: How often does your organization deploy code to production?

Why It Matters: Deployment frequency is a proxy for batch size. Small, frequent deployments mean:

  • Lower risk (less can go wrong)
  • Faster feedback (find problems sooner)
  • Faster time to market (features reach customers quickly)
  • Better team morale (see your work in production)

Performance Levels:

  • Elite: Multiple deployments per day (on-demand)
  • High: Between once per day and once per week
  • Medium: Between once per week and once per month
  • Low: Between once per month and once every six months

Example:

  • Low Performer: "We deploy every 2 months during maintenance windows"
  • Elite Performer: "We deploy 50+ times per day automatically"

How Fawkes Tracks It: Every ArgoCD sync to production is recorded as a deployment event.

2. ⏱️ Lead Time for Changes (LT)

Definition: How long does it take for a commit to go from version control to running in production?

Why It Matters: Lead time measures efficiency. Short lead times mean:

  • Faster feature delivery to customers
  • Quicker response to market changes
  • Reduced work-in-progress inventory
  • Higher developer satisfaction

Performance Levels:

  • Elite: Less than one hour
  • High: Between one day and one week
  • Medium: Between one month and six months
  • Low: More than six months

Example:

  • Low Performer: "I wrote this code 3 months ago. Still waiting for QA approval."
  • Elite Performer: "I committed code 20 minutes ago. It's already in production."

How Fawkes Tracks It: Measures time from Git commit to successful ArgoCD sync in production.

Important: This is NOT "time to write code." It's "time code sits waiting" in your process.

3. 🔧 Time to Restore Service (MTTR)

Definition: How long does it take to restore service when an incident occurs?

Why It Matters: MTTR measures resilience. Fast recovery means:

  • Less customer impact from incidents
  • Lower stress for on-call engineers
  • Better SLAs and reliability
  • Confidence to move fast (you can recover quickly)

Performance Levels:

  • Elite: Less than one hour
  • High: Less than one day
  • Medium: Between one day and one week
  • Low: More than one week

Example:

  • Low Performer: "Production is down. We need a 5-hour emergency change board meeting."
  • Elite Performer: "Production issue detected. Automatic rollback completed in 4 minutes."

How Fawkes Tracks It: Measures time from incident creation (Alertmanager) to resolution (successful deployment or rollback).

4. ❌ Change Failure Rate (CFR)

Definition: What percentage of deployments cause failures in production?

Why It Matters: CFR measures quality. Low failure rates mean:

  • Sustainable velocity (not breaking things constantly)
  • Lower operational burden
  • Better customer experience
  • More time for feature development (less firefighting)

Performance Levels:

  • Elite: 0-15%
  • High: 16-30%
  • Medium: 16-30%
  • Low: 16-30%

Note: 2023 research collapsed High/Medium/Low into same range. Elite performers stand out with <15%.

Example:

  • Low Performer: "Every Friday deployment requires weekend hotfixes."
  • Elite Performer: "We deploy 100 times per week with 5% failure rate."

How Fawkes Tracks It: Compares successful deployments to failed deployments (rollbacks, incidents within 24 hours of deploy).

Important: Some failure is expected and healthy! 0% might mean you're too risk-averse.

The Performance Spectrum

Here's how teams compare across the four metrics:

Performance Deployment Freq Lead Time MTTR Change Fail Rate
Elite On-demand (multiple/day) < 1 hour < 1 hour 0-15%
High 1/day - 1/week 1 day - 1 week < 1 day 16-30%
Medium 1/week - 1/month 1 week - 1 month 1 day - 1 week 16-30%
Low 1/month - 6/months 1 month - 6 months > 1 week 16-30%

Key Insight: Elite performers are 417x faster at deploying and 6,570x faster at going from commit to production than low performers!

Why These Four Metrics?

They Balance Speed and Stability

Speed Metrics:

  • Deployment Frequency
  • Lead Time for Changes

Stability Metrics:

  • Time to Restore Service
  • Change Failure Rate

You can't optimize for speed alone (you'll break everything) or stability alone (you'll move too slowly). Elite performers excel at all four simultaneously.

They're Predictive, Not Descriptive

These metrics don't just describe performance—they predict business outcomes:

  • Profitability: Teams with high DORA metrics are 2x more likely to exceed profitability targets
  • Market Share: 50% more likely to have higher market share
  • Productivity: 2x more likely to exceed productivity goals
  • Customer Satisfaction: 2x more likely to have happy customers

They Focus on Outcomes, Not Activities

Bad metrics: Lines of code written, hours worked, tickets closed Good metrics (DORA): How fast you deliver value and how reliably

The Business Case for DORA Metrics

Scenario: Legacy Bank vs. Digital Startup

Legacy Bank (Low Performer):

  • Deploys every 3 months
  • Lead time: 4 months from idea to production
  • MTTR: 3 days (requires emergency change approval)
  • CFR: 25% (1 in 4 releases has issues)

Impact:

  • New credit card feature takes 1 year to launch (competitors launch in 6 weeks)
  • When mobile app crashes, customers can't access accounts for 3 days
  • Developer turnover: 35% annually (frustration with slow process)

Digital Startup (Elite Performer):

  • Deploys 20x per day
  • Lead time: 2 hours from commit to production
  • MTTR: 15 minutes (automated rollback)
  • CFR: 8% (rigorous testing catches issues)

Impact:

  • New feature ideas tested with customers within days
  • Production incidents resolved in minutes, not days
  • Developer retention: 95% (engineers love working there)

Result: Startup captures 30% market share in 2 years despite having 1/100th the resources.

How Platform Engineering Improves DORA Metrics

A well-designed platform (like Fawkes) directly improves all four metrics:

Deployment Frequency ↑

  • Automation: CI/CD pipelines remove manual deployment steps
  • Self-Service: Teams deploy when ready, no waiting for tickets
  • Reduced Fear: Good testing and rollback make deployments safe

Lead Time ↓

  • Automated Testing: No waiting for manual QA
  • Fast Pipelines: Optimized builds complete in minutes
  • Simplified Process: Golden paths remove decision paralysis

MTTR ↓

  • Observability: Know immediately when things break
  • Quick Rollback: Automated rollback via GitOps
  • Runbooks: Standardized incident response

Change Failure Rate ↓

  • Quality Gates: Automated security scanning, testing
  • Consistent Patterns: Golden paths reduce errors
  • Progressive Delivery: Canary deployments catch issues early

The Platform Advantage: Manual processes hit scaling limits. Platforms enable teams to improve metrics continuously.

Common Misconceptions

❌ "We can't measure that in our organization"

Reality: If you deploy software, you can measure these metrics. Start simple with manual tracking if needed.

❌ "Our industry is different; this doesn't apply"

Reality: DORA research spans every industry from finance to gaming to healthcare. The metrics apply universally.

❌ "We need to slow down to improve quality"

Reality: Elite performers deploy MORE frequently AND have LOWER change failure rates. Speed and stability go together.

❌ "Our legacy systems prevent us from improving"

Reality: Legacy systems are a constraint, not an excuse. Many elite performers maintain legacy systems.

❌ "Leadership only cares about features, not metrics"

Reality: These metrics predict revenue, market share, and profitability. Leadership should care.

❌ "100% success rate is the goal"

Reality: Some failure is healthy. Elite performers have 8-15% CFR because they're taking appropriate risks.

How Fawkes Automates DORA Metrics

Fawkes collects DORA metrics automatically from your CI/CD pipeline:

Developer commits code
    ↓
Git webhook triggers Jenkins pipeline
    ↓ (Lead Time measurement starts)
Jenkins builds, tests, packages
    ↓
Artifact pushed to Harbor registry
    ↓
ArgoCD detects new image version
    ↓
ArgoCD syncs to Kubernetes (Deployment event recorded)
    ↓ (Lead Time measurement ends)
Prometheus records metrics
    ↓
Grafana dashboard updates in real-time
    ↓
Alertmanager detects any incidents
    ↓ (MTTR measurement if incident occurs)

Data Sources:

  • Git: Commit timestamps (lead time start)
  • Jenkins: Build results (quality signals)
  • ArgoCD: Deployment events (DF, lead time end, CFR)
  • Prometheus/Alertmanager: Incident detection and resolution (MTTR)

No Manual Work Required: Metrics update automatically with every deployment.

3. Demonstration (10 minutes)

📺 Video: Navigating Fawkes DORA Dashboards (10 minutes)

[VIDEO PLACEHOLDER] > See detailed script in supporting document

Key Takeaways from Demo

  1. Real-Time Updates: Metrics update with every deployment
  2. Multiple Views: Team-level, service-level, and organization-level dashboards
  3. Drill-Down Capability: Click any metric to see underlying data
  4. Trend Analysis: Compare current period to previous periods
  5. Actionable Insights: Dashboard highlights improvement opportunities

4. Hands-On Lab (20 minutes)

Lab Overview

You'll analyze DORA metrics for a sample application, identify performance bottlenecks, and make recommendations for improvement.

Time Estimate: 20 minutes Difficulty: Beginner Auto-Graded: Partially (calculations auto-checked; recommendations manually reviewed) Points: 60

Lab Environment

When you click "Start Lab", we'll provision:

  • ✅ Access to Grafana DORA dashboards
  • ✅ Sample data for 3 months (90 days)
  • ✅ 3 different teams with varying performance levels
  • ✅ Lab notebook for your analysis

Environment will be available for 24 hours from start time.

Lab Instructions

Part 1: Calculate Metrics (30 points)

You'll analyze "Team Alpha's" performance over the last 30 days.

Given Data (available in dashboard):

  • Total deployments to production: 45
  • Total commits: 180
  • Failed deployments (rollbacks): 7
  • Incidents reported: 3
  • Average time from commit to production: 6 hours

  • Average time to resolve incidents: 2 hours

  • Calculate Deployment Frequency (10 points)

Formula: Total deployments / Days in period

📝 Submit: What is Team Alpha's deployment frequency? (deployments per day)

✅ Validation: Auto-checked against correct calculation

  1. Calculate Lead Time for Changes (10 points)

Given: Average time from commit to production = 6 hours

📝 Submit: What is Team Alpha's lead time? Express in hours.

✅ Validation: Auto-checked

  1. Calculate Change Failure Rate (10 points)

Formula: (Failed deployments / Total deployments) × 100

📝 Submit: What is Team Alpha's change failure rate? Express as a percentage.

✅ Validation: Auto-checked against correct calculation

Part 2: Performance Classification (15 points)

  1. Classify Team Alpha's Performance (15 points)

Based on the metrics you calculated, classify Team Alpha according to DORA performance levels:

📝 Submit:

  • Deployment Frequency Level: [Elite/High/Medium/Low]
  • Lead Time Level: [Elite/High/Medium/Low]
  • Change Failure Rate Level: [Elite/High/Medium/Low]
  • Overall Classification: [Elite/High/Medium/Low]

✅ Validation: Auto-checked against DORA thresholds

Part 3: Compare Teams (15 points)

  1. Analyze Team Bravo vs. Team Charlie (15 points)

Open the "Team Comparison" dashboard and compare Team Bravo and Team Charlie.

Team Bravo:

  • DF: 0.3 per day (9 per month)
  • LT: 3 days
  • MTTR: 4 hours
  • CFR: 10%

Team Charlie:

  • DF: 2.5 per day (75 per month)
  • LT: 45 minutes
  • MTTR: 30 minutes
  • CFR: 18%

📝 Submit:

  • Which team is the higher performer overall? [Bravo/Charlie]
  • What is Team Charlie's biggest weakness? [DF/LT/MTTR/CFR]
  • If Team Bravo could improve one metric, which would have the biggest impact? [DF/LT/MTTR/CFR]
  • Explain your reasoning (2-3 sentences)

✅ Validation: Reasoning manually reviewed by instructors

Part 4: Identify Improvement Opportunities (Bonus)

  1. Recommend Improvements for Team Alpha (Bonus: +10 points)

Based on Team Alpha's metrics:

  • DF: 1.5 per day (High)
  • LT: 6 hours (Elite)
  • MTTR: 2 hours (Elite)
  • CFR: 15.6% (Elite)

📝 Submit:

  • Team Alpha is performing at Elite level across all metrics. However, what could they do to push even further? (3-5 specific recommendations)

Examples of good recommendations:

  • "Reduce deployment frequency variability (some days have 5 deploys, others have 0)"
  • "Investigate the 7 failed deployments to find common root causes"
  • "Implement chaos engineering to practice MTTR scenarios"

✅ Validation: Manually reviewed for thoughtfulness and actionability

Lab Submission

Once you've completed all tasks:

  1. Review your calculations in the lab notebook
  2. Ensure all required answers are recorded
  3. Click "Submit Lab" button

Grading:

  • Parts 1-2: Auto-graded immediately (45 points)
  • Parts 3-4: Reviewed within 24 hours by instructors (15 + 10 points)
  • Passing score: 48/60 (80%)

Troubleshooting Hints

Can't access Grafana?

  • Click "Open Grafana" from lab instructions
  • Use provided credentials (auto-populated)
  • Try incognito mode if having authentication issues

Calculations not matching?

  • Double-check your formulas
  • Ensure you're using correct time periods (30 days)
  • Round to 2 decimal places

Don't understand a metric?

  • Review the Theory & Concepts section
  • Check the DORA handbook link in resources
  • Ask in #dojo-white-belt on Mattermost

5. Knowledge Check (5 minutes)

Quiz: DORA Metrics Mastery

Instructions: Answer all 10 questions. You need 8/10 (80%) to pass. Unlimited attempts allowed.

Question 1

Which metric measures "how often" you deploy to production?

  • [x] A) Deployment Frequency
  • [ ] B) Lead Time for Changes
  • [ ] C) Mean Time to Restore
  • [ ] D) Change Failure Rate

Explanation: Deployment Frequency measures how often deployments occur.

Question 2

An elite performer's Lead Time for Changes is:

  • [x] A) Less than one hour
  • [ ] B) Between one day and one week
  • [ ] C) Less than one day
  • [ ] D) Between one hour and one day

Explanation: Elite performers have lead times less than one hour from commit to production.

Question 3

What does MTTR stand for?

  • [ ] A) Mean Time To Release
  • [ ] B) Mean Time To Recover
  • [x] C) Mean Time To Restore (Service)
  • [ ] D) Mean Time To Rollback

Explanation: MTTR is Mean Time To Restore Service—how long it takes to recover from incidents.

Question 4

Elite performers have a Change Failure Rate of:

  • [x] A) 0-15%
  • [ ] B) 16-30%
  • [ ] C) Less than 5%
  • [ ] D) 31-45%

Explanation: Elite performers maintain a CFR of 0-15%, significantly better than other performers.

Question 5

Which statement is TRUE about DORA metrics?

  • [ ] A) You must choose between speed (DF/LT) and stability (MTTR/CFR)
  • [x] B) Elite performers excel at all four metrics simultaneously
  • [ ] C) Only deployment frequency matters
  • [ ] D) These metrics only apply to startups, not enterprises

Explanation: Elite performers are fast AND stable—they excel at all four metrics at once.

Question 6

Your team deploys once per month. What performance level is this?

  • [ ] A) Elite
  • [ ] B) High
  • [x] C) Medium
  • [ ] D) Low

Explanation: Once per month is Medium performance (between once per week and once per month).

Question 7

Lead Time for Changes measures:

  • [ ] A) Time spent writing code
  • [x] B) Time from commit to production
  • [ ] C) Time in code review
  • [ ] D) Time spent in planning

Explanation: Lead time is commit to production—how long code waits in your process.

Question 8

Why do DORA metrics matter to business leaders?

  • [ ] A) They're required for compliance
  • [x] B) They predict profitability, market share, and customer satisfaction
  • [ ] C) They make engineers look good
  • [ ] D) They're easy to game

Explanation: DORA metrics are predictive of business outcomes—2x more likely to exceed profitability goals, etc.

Question 9

A team has 20 deployments and 5 failures in a month. What's their CFR?

  • [ ] A) 5%
  • [ ] B) 15%
  • [x] C) 25%
  • [ ] D) 50%

Explanation: CFR = (5 failures / 20 deploys) × 100 = 25%

Question 10

How does a platform like Fawkes improve DORA metrics?

  • [ ] A) By forcing teams to deploy more frequently
  • [ ] B) By hiding failure metrics
  • [x] C) By automating pipelines, testing, and providing fast feedback
  • [ ] D) By reducing the number of engineers needed

Explanation: Platforms improve metrics through automation, quality gates, and fast feedback loops—making the right things easy.

Quiz Results

Score: X / 10

  • ✅ Passed (8+): Excellent! You understand DORA metrics deeply.
  • ❌ Not Yet (<8): Review the content and try again.

6. Reflection & Next Steps (5 minutes)

What You Learned

Congratulations! 🎉 You've completed Module 2. Let's recap:

✅ You now understand:

  • The Four Key Metrics and what they measure
  • Why these metrics predict business success
  • How to calculate and interpret DORA metrics
  • The difference between Elite and Low performers
  • How Fawkes automates metrics collection

✅ You can now:

  • Analyze DORA dashboards and spot issues
  • Make data-driven recommendations for improvement
  • Explain metrics to business stakeholders
  • Use metrics to prioritize platform improvements

How This Connects to Your Work

For Developers:

  • You understand what "good" looks like (Elite benchmarks)
  • You can advocate for improvements using data
  • You know how to track your team's progress

For Platform Engineers:

  • You can measure platform impact objectively
  • You know which improvements matter most
  • You can demonstrate ROI to leadership

For Leaders:

  • You have a data-driven framework for investment decisions
  • You can benchmark against industry standards
  • You can track improvement over time

Real-World Application Exercise

This Week, Try This:

  1. Measure Your Current State

  2. Track deployments for one week

  3. Calculate your team's current DORA metrics

  4. Be honest—no judgment, just data

  5. Identify One Improvement

  6. Pick the metric with the most room for improvement

  7. Brainstorm 3 concrete actions to improve it

  8. Estimate impact and effort

  9. Share Your Findings

  10. Present current state to your team (5 min standup)
  11. Discuss: "What's our biggest bottleneck?"
  12. Agree on one improvement to try

Reflection Questions

Take 2 minutes to think about:

  1. Which metric surprised you most?

  2. Did your team's performance match your intuition?

  3. What's your team's biggest opportunity?

  4. Which metric, if improved, would have the most impact?

  5. What's blocking improvement?

  6. Technical debt? Process issues? Cultural resistance?

  7. Who needs to know this?

  8. Which leader should see your team's DORA metrics?

Additional Resources

📚 Further Reading:

🎥 Videos to Watch:

  • "DORA Metrics Explained" by Dr. Nicole Forsgren (15 min)
  • "Why DORA Metrics Matter" by Gene Kim (20 min)
  • "Implementing DORA Metrics" by Charity Majors (30 min)

🛠️ Tools:

  • Four Keys Project - Open source DORA metrics tool
  • Sleuth - Commercial DORA tracking (Fawkes alternative)
  • LinearB - Engineering intelligence platform

💬 Community:

  • Share your team's metrics (anonymously!) in #dojo-metrics
  • Join the DORA community discussions
  • Help others interpret their data

Preview: Module 3

Next Up: GitOps Principles

In Module 3, you'll learn:

  • What GitOps is and why it's transforming deployments
  • Declarative infrastructure and desired state
  • How ArgoCD implements GitOps
  • Pull-based vs. push-based deployments
  • Making your first GitOps change

Time: 60 minutes Hands-On: Make a GitOps deployment using ArgoCD

Get Ready: Think about how your team currently deploys applications. Who has access? How is it documented? What could go wrong?

Module Completion

✅ You've Completed Module 2

Next Steps:

  1. ✅ Mark this module complete in your Backstage profile
  2. 📊 View your progress on the Dojo dashboard
  3. 💬 Share your DORA metrics insights in #dojo-achievements
  4. ➡️ Continue to Module 3 when ready

Time Investment: 60 minutes Skills Gained: DORA metrics analysis, performance benchmarking Progress: 2 of 4 modules toward White Belt (50% complete)

Questions or Issues?

  • 💬 Ask in #dojo-white-belt on Mattermost
  • 📧 Email: dojo@fawkes.io
  • 🐛 Report bugs: GitHub Issues

Feedback?

  • Rate this module (takes 30 seconds)
  • What worked well? What could be better?
  • Help us improve the learning experience!

Module Author: Fawkes Learning Team Last Updated: October 2025 Version: 1.0 Based On: DORA State of DevOps 2023 Report