Testing In Production
summary
Claims that running tests against live systems raises confidence and velocity by catching real-data bugs before release and lowering fix time, with concrete practices: use feature flagging (Split or LaunchDarkly) to target developers and bots, create identifiable test users or backend test flags so analytics and third parties ignore robot actions, schedule automated checks (Robot Framework / Jest / Protractor) with Jenkins/CircleCI, alert via PagerDuty/Slack, and roll changes via a production canary so failures hit few users and can be rolled back immediately.
AI-written summary · about
chapters (15)
- 0:00Start
- 4:20What's wrong with staging environments?
- 7:07Testing in production only works when...
- 12:32Parallel Feature Flags
- 17:23Job Scheduler
- 19:16Tools
- 20:12Solution
- 23:32Setup & Teardown
- 25:18Why use a production canary?
- 26:21Canary Setup
- 27:18Risk Mitigation
- 28:58Performance Testing in Production
- 30:28Testing in a Containerized Environment
- 31:47Where do prod tests operate?
- 32:22Shifting your company's testing culture
