2 Weeks is Too Long - On Demand Feature Deployments with Drupal
summary
Argues that two-week release cadence is unnecessary and shows concrete changes: deploy code daily behind feature flags, run isolated PR previews with Pantheon multidev, automate checks with Cypress, and extract marketing forms into a Formik microservice. A client ships ordering API code to production three weeks before go-live and flips the flag during a 20-minute window, avoiding a 16-hour blackout. Uses Drupal config management and cleanup of toggles afterward.
AI-written summary · about
chapters (17)
- 0:00Introduction and goals
- 1:47Sprints and releases history
- 5:22Why teams run long release cycles
- 7:56Mindsets and inflexible processes
- 11:11Breaking the two-week cycle mindset
- 13:39Feature-based environments and automation
- 16:52Keys to moving faster
- 17:13What are feature flags
- 19:42Drupal feature flag examples
- 22:11Microservices and decoupling
- 23:39Success story and benefits
- 25:48Drupal landscape and tools
- 27:58Q&A on feature flag cleanup
- 28:41Q&A on config issues
- 30:06Q&A on project types and team alignment
- 32:14Q&A on ticket planning and features
- 33:16Closing and final thoughts
