AWS Solutions Architect Associate

Correct — You design across AZs within a Region for high availability.

Correct — Spreading resources across AZs survives a single-AZ failure.

Correct — Multi-AZ is the core HA pattern on AWS.

Correct — Root has unrestricted power and few actions truly require it.

Correct — Default-deny, explicit-deny-wins is the core model.

Correct — Roles deliver short-lived credentials — no long-lived keys.

Correct — The instance gets auto-rotated temporary credentials via the metadata service.

Correct — Scope actions and resources tightly.

Correct — Manage permissions by group rather than per user.

Correct — Even an account admin cannot exceed an SCP.

Correct — AWS runs the infrastructure; you secure your workloads/config/data.

Correct — Isolation per account, one payer, volume discounts pooled.

Correct — You choose instance type, AMI, and manage the OS.

Correct — You pay per invocation/duration; AWS runs the compute.

Correct — Fargate runs containers serverlessly — no EC2 to manage.

Correct — Commit to usage for a significant discount.

Correct — They are cheap but can be reclaimed with short notice.

Correct — 11 nines of durability; ideal for files, backups, static assets.

Correct — Lower storage cost, retrieval fee, still instant access.

Correct — Like a virtual disk; snapshot to S3 for durability/backup.

Correct — A managed, elastic NFS shared across instances and AZs.

Correct — Fast local disk for scratch/cache, not persistence.

Correct — Bake a golden AMI for consistent, fast launches.

Correct — Transition to IA/Glacier or expire objects on a schedule.

Correct — You define subnets, routing, and gateways within it.

Correct — Public subnets route to the IGW; private ones do not.

Correct — Outbound-only internet for private subnets.

Correct — You allow inbound/outbound; replies flow back automatically.

Correct — Latency, failover, weighted, and geolocation routing.

Correct — Lower latency and offload for static/dynamic content.

Correct — Content-based routing to target groups.

Correct — L4, extreme performance, static IPs.

Correct — You get a managed engine; AWS runs the undifferentiated ops.

Correct — A standby in another AZ takes over on failure.

Correct — Scales seamlessly; no servers to manage.

Correct — Offload hot reads to a managed in-memory cache.

Correct — Elasticity plus self-healing across AZs.

Correct — The classic resilient web-tier pattern.

Correct — The queue buffers work; consumers process at their own rate.

Correct — One publish reaches queues, functions, HTTP endpoints, etc.

Correct — Cheaper than warm/hot standby, faster than backup-restore.

Correct — Cost and recovery speed both increase along this path.

Correct — How long you can be down; drives the DR strategy.

Correct — Keep state in a shared store (DynamoDB/ElastiCache), not on the instance.

Correct — Higher performance and availability, MySQL/Postgres-compatible.

Correct — Event-driven, loosely-coupled integration at scale.

Correct — Offload read queries to replicas; the primary handles writes.

Correct — Everything fails eventually; architect to survive it.

Correct — KMS integrates with these services for at-rest encryption.

Correct — A web application firewall filtering L7 attacks.

Correct — Standard is automatic; Advanced adds more.

Correct — Managed, rotated, access-controlled, nothing in code.

Correct — Commit for the predictable base; flex the rest.

Correct — Use CloudWatch/Compute Optimizer to find the right size.

Correct — Pay less for data you rarely touch.

Correct — The six pillars for reviewing an architecture.

Correct — Design to withstand and recover from disruptions.

Correct — Managed detection plus a central findings view.

Correct — Monitor resources and trigger actions on thresholds.

Correct — Visibility plus deliberate purchasing decisions.