Applying SRE Principles to CI/CD
summary
Applies SRE practices to CI/CD with concrete metrics and tools: measures show Buildkite users cumulatively spend 9,413 days retrying builds in one month, uses SLIs/SLOs/error budgets (examples: start within 30s with an error budget of 33 builds/4 weeks; total run time <5 min with 33 breaches/4 weeks; suite reliability >87% with 77 low-score runs). Uses Buildkite Test Analytics, OpenTelemetry, Datadog/Honeycomb/CloudWatch, autoscaling agents and parallel steps; prototype finds 100 flaky tests and speeds a test from 35s to under 3s.
AI-written summary · about
chapters (8)
- 0:00Introduction and background
- 2:55The daily struggle of CI/CD
- 5:47The impact of flaky tests
- 11:43Applying SRE principles
- 16:53Defining SLIs, SLOs, and budgets
- 22:52Practical implementation
- 28:27Managing technical debt
- 31:45Q&A session
related talks
27mApplying SRE Principles to CI/CD
28mKeeping Developers Happy with a Fast CI
47mHow to win in software development with tighter feedback loops
27mContinuous Reliability. How?
36mThinking like a site reliability engineer to improve CI
24mCI/CD Best Practices to Deliver Software Faster
33mHigh performing engineering teams and the Holy Grail
32mPractical Tips & Tricks for CI/CD Success