Skip to main content

Measuring Metrics

Updated Mar 17, 2021 ·

Metrics​

Metrics assess the effectiveness of DevOps practices on pipeline performance and software quality.

  • Focus on efficient, high-quality software development, deployment, and operation
  • Metrics provide critical insights into overall pipeline performance

Deployment Frequency​

Reflects pipeline health.

  • Indicates development activity
  • Ideally trending upward during active development
  • Stabilizes as software reaches maturity

Mean Time to Recovery (MTTR)​

Tracks average time to resolve production issues.

  • Measures time to recover from outages
  • Time resolving bugs, and security issues affecting users
  • Goal is rapid recovery to maintain user experience

Failures are inevitable. Creating flexible software and infrastructure helps minimize impact and improve recovery speed, aiming to reduce failure frequency.

Mean Time to Discovery (MTTD)​

Measures how quickly failures are detected.

  • Start point is when an issue is introduced in production, not when it’s discovered
  • Detection method could be automated or user-reported, affecting response times

System Availability​

Tracks system uptime and availability of each software component.

  • Monitors uptime for components like load balancers, servers, and CDNs
  • Low uptime indicates areas needing improvement (e.g., if web servers are only 90% up monthly)

Service Performance​

Identifies performance issues through response times and load times.

  • Performance metrics include API response and webpage load times, crucial for optimizing user experience
  • Track changes after deployments to pinpoint performance-related code issues

Customer Complaints​

Analyzes customer feedback on issues.

  • Track trends to identify frequent issues and improve preventive measures
  • Preventive measures reduce recurring problems and support a proactive improvement culture

Lead Time​

Tracks time from feature request to release.

  • Goal is rapid delivery without sacrificing quality to maintain efficiency
  • Fast turnaround supports experimentation, preventing software stagnation

Conclusion​

These metrics provide insights to enhance the pipeline. They should inform improvements, not impose arbitrary goals.


Feedback