Testing in Production
summary
Argues for running targeted tests inside production by deploying code behind feature flags and rolling out with canary releases (percentage rollout) while separating test data via an attribute like is_test=true. Recommends automating test users, using a job scheduler (CircleCI/Jenkins), alerting (PagerDuty + Slack), and tools such as Split or LaunchDarkly and Robot Framework for test automation. Starts small (CSS or trivial flow), uses AA tests to validate flags, and reports faster releases, fewer post‑deploy bugs, and higher release confidence.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:45DATA MISMATCH
- 11:14PRODUCTION
- 13:22How do you know what to test in prod?
- 14:34Recommended Tools
- 16:16Risk Mitigation
- 18:55Think of the last feature your team deployed
- 19:39Culture Shift
