Testing in Production
summary
Argues for running critical end to end checks in live traffic by deploying behind feature flags and rolling out with canary releases, using targeted test accounts or an override abstraction to fake flag states, separating test and real data in dashboards, and automating runs with a job scheduler and alerts. Recommends Split for flags and Robot Framework for automation, claims faster, more accurate tests and higher release confidence when guarded by isolation and incremental rollouts.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:45DATA MISMATCH
- 11:20PRODUCTION
- 13:45How do you know what to test in prod?
- 15:38Recommended Tools
- 17:44Risk Mitigation
- 19:20Think of the last feature your team deployed
- 19:51Culture Shift
