Question

Your product has been in production for 18 months using continuous deployment with automated testing. The development team releases small increments to production multiple times per week. Recently, the operations team escalated concerns that production incidents have increased 40% over the past two months, with several customer-impacting outages. The VP of Engineering questions whether the continuous deployment approach is too risky and suggests returning to monthly release cycles with more comprehensive manual testing. Your analysis shows that while incident count increased, 85% of incidents are caught and auto-rolled back within 2 minutes before customers are affected. The remaining 15% that impact customers are primarily due to third-party API changes and infrastructure issues, not code defects. The team's automated test coverage is 87% and their mean time to recovery is 8 minutes. What should you recommend?

Accepted Answer

Maintain the continuous deployment approach but implement enhanced monitoring and alerting for third-party dependencies, and conduct failure mode analysis on the customer-impacting incidents

Answer

Implement a hybrid approach with automated deployments for minor changes but require manual approval and testing gates for any changes touching critical customer-facing features

Answer

Transition to a weekly release cadence as a compromise that maintains reasonable agility while allowing more comprehensive pre-production validation

Answer

Keep continuous deployment but add a mandatory 24-hour soak period in a production-equivalent staging environment before any code reaches actual customers

PMP Practice Question: Process — Agile, Hard

More Process Questions