Continuous Integration at Haptik : Our Journey to Hourly Releases
summary
Claims a move from monthly releases to hourly releases by wiring Jenkins, AMIs and AWS spot instances into a blue/green rollout, running unit and integration tests with Selenium and enforcing PR checks. It reports shipping several times a day, managing 40–50 bots and 1.2–1.3M messages/day, and building a custom bot tester that validates conversation endpoints rather than exact copy. Emphasis falls on test-driven culture, automated data/config propagation, and making developers own deploys end-to-end.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 0:20About Haptik
- 1:08Team Structure
- 2:22Tech Stack
- 2:44Problems
- 4:10Goals
- 6:10Environments
- 8:18Testing
- 10:17Data
- 12:00Key takeaways
- 15:05Building a feature from scratch
- 23:04App testing
- 23:52Testing tool
- 26:23CI pipeline
- 27:15Focus
related talks
32mContinuous Integration at Haptik: our journey to hourly releases
35mMassively Continuous Integration: From 3 days to 30 minutes
35mFrom 6 Month Waterfall To 1h Code Deploys
32mBuilding Continuous Integration and Feedback
29mOur Journey to Fully Automated Weekly Releases
53mTransforming testing methodologies: From manual to robust strategies
27mAutomation, Innovation and Continuous Delivery
26mHuman Trust in Continuos Deployment