Free SAP-C02 practice test: 30 questions with answers

Multi-AZ synchronous replication keeps a standby in another zone for automatic failover with minimal data loss. Single-AZ instances, same-AZ replicas and laptops do not survive zone failures.

Lambda event source mapping with a DLQ and bounded retries captures repeated failures for analysis. Polling every second, local file storage and silent deletion do not meet the requirement.

ECS on Fargate removes server management, and service auto scaling and an ALB handle scale and routing. Manual Docker, Lambda for long-running services and colocation servers do not fit.

DynamoDB with DAX delivers very high throughput and microsecond-to-millisecond reads for key-based access. Redshift is analytical, Neptune is a graph database, and Glacier is archival.

Event-driven triggers start container work on demand and stop when done, which scales to zero. Permanent polling, laptop cron and manual processing do not fit the requirement.

PrivateLink provides private connectivity to services through interface endpoints. Public gateways, home-office VPNs and NAT gateways still use public routing.

FSx for Lustre provides high-throughput POSIX-compatible shared storage for HPC workloads. S3 has no POSIX mount by default, Glacier is archival, and Snowcone is for edge data collection.

Global Accelerator provides static anycast IP addresses and routes to the nearest healthy endpoints. Simple DNS routing does not provide static IPs, changing origins break static address expectations, and VPNs are not for public clients.

Cost Explorer with tags and anomaly alerts identifies drivers and unusual spikes. Yearly invoice reviews, disabled monitoring and guessing do not provide timely insight.

Commitments for baseline and elastic capacity for peaks align cost with usage. Peak-all-day, one large instance and disabled scaling waste money or reduce resilience.

CloudWatch alarms and cross-account observability give timely visibility, and SNS delivers notifications. Daily manual checks, ignoring errors and disabled metrics delay response.

AWS Config rules continuously evaluate security group configurations against policy. The other services are unrelated to network configuration checks.

Regular, timed tests in isolation prove recovery and reveal gaps without impacting production. Assumptions, outage-only testing and skipping tests leave risk unverified.

Config rules with automatic remediation fix public buckets quickly. Quarterly manual steps, reminder emails and turning off S3 do not provide timely automated remediation.

Spot and diversified Auto Scaling suit interruption-tolerant work and lower cost. Critical databases on Spot and All-On-Demand batch approaches do not reduce cost for that type of workload.

Performance Insights shows database load by wait events and SQL statements. Budgets, Macie and Shield do not analyze database performance.

An organization trail to a dedicated archive account centralizes logs and keeps them outside application-team control. Local buckets, disabled trails and shared root credentials fail the requirement.

IAM Identity Center federates identities and assigns permission sets across accounts with temporary credentials. Permanent keys, shared root logins and anonymous access are not appropriate.

Transit Gateway shared with RAM lets accounts attach VPCs to a central hub without full-mesh peering. Peering each pair scales poorly, public IPs do not provide routing, and NAT instances are not a hub.

SCPs are preventive guardrails that cap permissions across accounts. Written policies, after-the-fact rules and monthly reviews do not stop use.

A scoped bucket policy grants precise cross-account read access to one prefix. Root credentials, public access and manual copies do not meet least privilege.

Regions in the permitted country and replication within allowed boundaries satisfy residency and latency. Global replication violates residency, CEO proximity is irrelevant, and ignoring residency breaks compliance.

Consolidated billing combines charges and SCPs apply to member accounts regardless of local administrator rights. Organizations can hold many accounts, and management account resources are billed.

The strangler fig pattern replaces functionality incrementally behind a routing layer. Big-bang cutovers, unchanged lift-and-shift and hard-coded client routing do not describe incremental modernization.

Drills and measured results validate recovery objectives. Skipping drills and relying on untested backups leave recovery unproven.

RDS for SQL Server preserves the engine, so the application changes little. A graph rewrite, laptops and flat files do not fit the requirement.

DataSync performs incremental transfers with bandwidth control, which suits constrained links until cutover. Snowmobile is for very large one-time transfers, and email and paper are impractical.

SNS fan-out to SQS queues or Lambda decouples services and supports multiple consumers. Polling a shared table, point-to-point calls without retries and email between services are fragile designs.

Data validation and application testing confirm that the target is correct and ready. Deleting backups, secrecy and disabling monitoring increase risk.

DataSync transfers and synchronizes files with bandwidth controls. Snowcone is a device, Glacier Select queries archives, and Step Functions orchestrates workflows.