Cloud and Infrastructure Operations
Operate for failure, cost, and evidence
Cloud and infrastructure operations covers service models, resilience, observability, automation, and cost. Know IaaS, PaaS, and SaaS, and the shared responsibility model: the provider secures the cloud; the customer secures what they put in it. Responsibility shifts with the service model-patching an OS is yours on IaaS, the provider's on PaaS. Containers package an app and dependencies; Kubernetes orchestrates containers; serverless abstracts servers but still has limits. Infrastructure as code makes environments reproducible; configuration management keeps them consistent.
Resilience terms are commonly tested: high availability reduces downtime within a region; disaster recovery restores service after a major loss; fault tolerance continues despite component failure; backup is a recoverable copy. Replication is not backup-it copies deletions too. Observability uses logs, metrics, and traces; SLIs measure service behavior, SLOs set targets, and SLAs are contractual. Error budgets make reliability trade-offs explicit.
Traps: assuming the provider patches everything; treating multi-region as automatic DR without failover testing; using autoscaling to mask inefficient code; and forgetting egress and idle-resource costs. Another trap is quoting four nines without defining the measurement window and exclusions.
Worked example: a 99.9% monthly SLO allows about 43 minutes of downtime. If the service has already consumed 30 minutes by day 20, release risk should be reduced until the budget recovers.
Rule of thumb: design for failure, test restore and failover, and track unit cost alongside uptime.