Change Failure Rate Calculator
Calculate the DORA Change Failure Rate metric from total and failed deployments, and see your DevOps performance tier.
Inputs
Total number of deployments/changes to production in the measurement period.
Number of those deployments that resulted in a failure (rollback, hotfix, incident, degraded service).
Change Failure Rate
6.67%
DORA Performance Tier
High
Successful Deployments
112
Tier Threshold
High: 5% – <10%
Step by step
Change Failure Rate: failed / total × 100
8 / 120 × 100
= 6.67%
Successful deployments
120 − 8
= 112
DORA performance tier
High: 5% – <10%
= High
How it works
Change Failure Rate (CFR) is one of the four key DORA metrics, measuring the percentage of deployments to production that result in a degraded service and require remediation (rollback, hotfix, patch). Formula: CFR = failed_deployments / total_deployments × 100. DORA's research-backed benchmarks classify teams as Elite (<5%), High (<10%), Medium (<15%) or Low (≥15%) performers based on this rate.
Formula
change_failure_rate = (failed_deployments / total_deployments) × 100
- failed_deployments
- Deployments that caused a failure requiring remediation
- total_deployments
- Total production deployments in the period
Frequently Asked Questions
What counts as a 'failed' deployment for CFR?
Any deployment that causes a service degradation requiring remediation — this includes rollbacks, hotfixes, patches, or incidents directly attributable to the change, not just deployments that technically failed to complete.
Why is a lower CFR better even at the cost of shipping less often?
DORA's research shows elite performers achieve both high deployment frequency AND low change failure rate simultaneously — the two aren't a tradeoff for high-performing teams, since better automated testing, smaller batch sizes and progressive delivery reduce failures while increasing speed.
How is CFR different from a bug count or defect rate?
CFR is scoped specifically to changes deployed to production that cause a production incident or require remediation, not all bugs found in QA or all defects reported by users — it's a delivery-pipeline health metric, not a general quality metric.
What's a good sample size for measuring CFR?
DORA benchmarks are typically measured over a rolling window of weeks to months with enough deployments (dozens or more) to avoid a single incident skewing the percentage disproportionately.
You might also need
- Platform Reliability Score CalculatorCommonly used together
- Engineering Productivity CalculatorCommonly used together
- SLI CalculatorCommonly used together
- Mean Time to Recovery MTTR CalculatorAlso in DevOps Metrics & DORA
- Deployment Frequency CalculatorAlso in DevOps Metrics & DORA
- Lead Time for Changes CalculatorAlso in DevOps Metrics & DORA