liveBook · Manning

Change failure rate

As explained in Grokking Software Architecture

Change failure rate is the proportion of software changes that result in a failure requiring remediation or rollback.

Change failure rate is the percentage of deployments that degrade production and require immediate intervention. It measures operational stability, not delivery speed. section 13.8

Read it alongside deployment frequency, lead time for changes, and time to restore service. Together, these metrics offer a diagnostic signal about how architecture performs under production pressure. section 13.8 section 13.12

Pipeline safety has a visible tradeoff:

  • Automated fitness functions in commit gates can catch regressions before deployment; high failure rates point to poor system testability. section 13.8
  • Skipping automated validation to ship faster means accepting higher change failure rates and longer recovery times. section 13.9.4

Making the risk explicit

When leaders want to bypass deployment checks for a deadline, the Knight Capital case can frame the risk. An Architecture Decision Record can document the choice and the risk the business accepts. section 13.9.4

Figure 13.12 The Dual Axes of DORA Metrics. High-velocity development metrics (Deployment Frequency and Lead Time) are continuously counterbalanced by operational stability metrics (Change Failure Rate and Time to Restore Service).

The figure places change failure rate on the stability axis opposite deployment frequency and lead time on the delivery-speed axis. section 13.8

How Grokking Software Architecture covers it

How deep it goes
Defines the rate as deployments degrading production and requiring intervention, then links high rates to poor testability.
Examples
Uses bypassed pipeline validation and the Knight Capital case to frame the business risk of skipping deployment safety.
What you can do after
Evaluate architecture by reading change failure rate alongside delivery speed and recovery time.
How it connects
Connects deployment failures with automated fitness functions, commit gates, deployment frequency, lead time, and time to restore service.

Also explained in

Compare how 5 books explain change failure rate →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Grokking Software Architecture. The section links point to the parts of the book it draws on.