
You are responsible for deploying a critical production service that supports customer-facing traffic. The team wants to reduce release risk while still shipping frequently, and you need an approach that balances speed, safety, and operational clarity.
What strategies do you use for rolling updates, canary releases, and ensuring zero-downtime deployments? How do you decide between them, define success criteria, and prepare rollback plans if the release degrades production?