Cloud Deployment and Operations
What to know
Operations covers how you build, release, and run workloads. Know Infrastructure as Code (IaC) - declarative templates that make environments reproducible and reviewable - and CI/CD pipelines that automate build, test, and deploy. Understand immutable infrastructure (replace, don't patch in place) and blue/green and canary release patterns for low-risk change. Observability rests on metrics, logs, and traces; alerting should key off symptoms (latency, error rate, saturation) rather than causes alone. Know RTO (time to restore) and RPO (acceptable data loss) for backup and DR planning, plus backup verification and restore testing.
Traps
Questions conflate backup with replication: replication copies corruption and deletions, so it is not a backup. IaC drift - manual changes diverging from templates - is a favorite. Don't confuse rolling updates (gradual replacement, mixed versions during rollout) with blue/green (two full environments, instant switch and rollback). Also separate deployment strategy from scaling strategy.
Worked example
A team wants near-zero-downtime releases with instant rollback. Answer: blue/green - stand up the new version, shift traffic, keep the old environment warm for rollback. If the requirement were gradual risk reduction with real-user traffic, canary fits better. Rule of thumb: choose the release pattern by the rollback requirement, not by how fancy it sounds.
Rule of thumb
Automate everything twice - once to create it (IaC) and once to prove it (tests, restore drills).