Cloud Operations and Optimization
Core concepts
Operations questions cover monitoring, cost, and performance tuning. Know CloudMonitor for metrics and alarms, ActionTrail for API audit, and Log Service for centralized logging. Understand auto scaling policies (target tracking, scheduled, step) and how they interact with load balancers and health checks.
Cost optimization
Recognize levers: right-sizing instances, reserved instances and savings plans for steady workloads, spot instances for fault-tolerant batch work, lifecycle rules and storage-class tiering for OSS, and releasing idle resources. Know that pay-as-you-go suits spiky loads while commitments suit predictable baselines.
Reliability operations
Understand health checks, graceful scaling, and the difference between reactive alarms and proactive capacity planning. Familiarize yourself with backup and snapshot policies and how RPO drives backup frequency.
Traps
Questions often pair a cost goal with a performance SLA - the cheapest option may violate latency or availability. Watch for spot instances proposed for stateful or latency-critical services. Another trap: alarms without action, or scaling policies that ignore warm-up time and cause thrashing.
Rule of thumb
Optimize in order: eliminate waste (idle resources), then right-size, then commit (reserved/spot), then tune architecture. Never trade away a stated SLA for cost savings unless the question explicitly prioritizes cost.