Continuous Performance Testing in DevOps
summary
Claims that traditional end‑of‑cycle performance testing is dead and argues for a two‑year migration to continuous performance testing that actually works. Implementation uses unit/component performance tests (only ~3% of sampled projects had them), CI triggers, load generators plus APM feeding a time‑series DB, synthetic transactions in production, and a virtual CoE and performance owner role. A surprising data point: only 28% of elite DevOps teams automate perf tests. Started small with a few component tests, iterated, and achieved sustainable frequent feedback.
AI-written summary · about
chapters (17)
- 0:00Continuous performance testing intro
- 4:40Traditional testing fails in DevOps
- 7:09What is continuous testing
- 9:15Missing pieces in DevOps culture
- 10:20Defining performance factors
- 13:56Building accountability
- 15:46Team skills and tooling
- 17:37Performance in each pipeline stage
- 19:23Unit testing at commit
- 21:08Integration testing stage
- 23:56Component testing monoliths
- 26:24Acceptance performance evaluation
- 28:51Continuous testing technologies
- 31:20Implementation lessons
- 35:10Supporting factors
- 36:58Safe production testing
- 40:10Key takeaways
related talks
41mContinuous Performance Testing in DevOps
41mPerformance Engineering in the Age of DevOps and Agile
33mPerformance Testing in Agile and DevOps
34mWhy continuous testing is essential
29mProgressive Testing to Meet the Performance Imperative
44mStrategy for Acing Continuous Testing in DevOps
37mThe Evolution of Performance Testing to Performance TestOps
42mTesting in a Continuous Delivery World