AMAADOR ACADEMYCertificatsVérifier un certificat

SOAC03

Test final — Préparation à l’examen AWS Certified CloudOps Engineer - Associate (SOA-C03)

Répondez aux 50 questions. Il faut 80 % ou plus pour réussir. Vous pouvez réessayer autant de fois que vous le souhaitez.

Ce test final est disponible en anglais.

Suivre la formation →

Question 1An application already ships its logs to a CloudWatch Logs log group. The team wants an alarm when more than five lines containing "OutOfMemoryError" appear in five minutes. What should you configure?
Question 2Which statement about SOA-C03 results is correct according to the AWS exam guide?
Question 3On-call engineers are paged several times a night by a latency alarm during harmless spikes. They want a page only when high latency and a high rate of HTTP 5xx errors happen at the same time. What is the best solution?
Question 4You configure an S3 bucket to send object-created events to an SNS topic. The console returns "Unable to validate the following destination configurations". What fixes it?
Question 5An application publishes a custom metric QueueDepth with the dimension InstanceId. You create an alarm on QueueDepth with the dimension AutoScalingGroupName. The alarm stays in INSUFFICIENT_DATA although the metric graph shows data. Why?
Question 6Which requirement needs CloudTrail configuration beyond the default event history?
Question 7A central operations team must watch EC2, RDS and Lambda metrics from 12 accounts in three Regions on one screen without signing in to each account. What should they set up?
Question 8A database on a 100 GiB gp2 volume performs well for about 30 minutes of heavy load, then slows sharply. CloudWatch shows BurstBalance falling to zero. What is the most cost-effective fix without downtime?
Question 9During traffic spikes, hundreds of concurrent Lambda functions connect to an RDS for MySQL instance and the database returns "Too many connections". CPU on the database stays low. What should you do?
Question 10A tightly coupled HPC simulation runs on 16 instances that exchange messages constantly. Engineers need the lowest network latency and highest packets per second between nodes. How should the instances be launched?
Question 11Field teams in Southeast Asia upload 2 GB video files to an S3 bucket in us-east-1. Uploads are slow and often fail near the end, so they must restart. What improves both speed and reliability?
Question 12An EventBridge rule created with the AWS CLI should invoke a Lambda function. The rule's MatchedEvents metric rises, but FailedInvocations rises too and the function never runs. What is the likely cause?
Question 13Linux web servers in three AZs need a shared POSIX file system. Most files are rarely read after 30 days. The team wants low management effort and lower storage cost for old files. What should you choose?
Question 14Which AWS service recommends better-sized EC2 instance types, EBS volumes and Lambda memory settings from historical utilisation?
Question 15An Auto Scaling group serves an API whose load changes unpredictably during the day. The team wants average CPU kept near 50% with the least ongoing tuning. Which scaling policy fits best?
Question 16A new Application Load Balancer returns HTTP 503 to every request. The listener forwards to a target group, and the Auto Scaling group shows four healthy instances, but the target group lists no targets. What should you do?
Question 17A failover record must switch DNS to a standby Region when an internal service in private subnets becomes unhealthy. The service has no public endpoint. Which Route 53 health check works?
Question 18Which statement about an RDS Multi-AZ DB instance deployment is correct?
Question 19A DynamoDB table in provisioned mode is throttled during sudden, unpredictable marketing spikes. The team does not want to forecast capacity. What should you change?
Question 20A product catalogue on RDS receives the same read queries thousands of times per minute and database CPU is near 90%. Data changes a few times per hour. What reduces database load most?
Question 21Every morning at 08:00 traffic jumps. New instances in the Auto Scaling group need five minutes to boot and load configuration, so users see slow responses until they are ready. Which change helps most?
Question 22Across 30 accounts, every EBS volume, RDS instance and DynamoDB table tagged backup=daily must be backed up daily, kept 35 days and copied to a second Region. New accounts must be covered automatically. What should you use?
Question 23At 14:05 a developer ran a DELETE without a WHERE clause on an RDS for PostgreSQL instance. Automated backups are kept for 7 days. You need the data as it was at 14:04. What should you do?
Question 24A web application with an RDS database needs a disaster recovery plan for a Regional outage with an RPO of 15 minutes and an RTO of about one hour, at the lowest cost that meets both. What should you choose?
Question 25A regulator requires that backup recovery points cannot be deleted or shortened by anyone, including the root user, for seven years. Which feature meets this?
Question 26You try to share EBS snapshots encrypted with the AWS managed key aws/ebs with a disaster recovery account, but the DR account cannot use them. What must you do?
Question 27You add a Cross-Region Replication rule to a versioned bucket that already holds 10 million objects. New uploads replicate, but the existing objects never appear in the destination. What should you do?
Question 28On an Aurora MySQL-compatible cluster with Backtrack enabled, a faulty migration corrupted a table 10 minutes ago. The team wants to rewind quickly without creating a new cluster. What should they do?
Question 29A CloudFormation stack failed on its first creation and now shows ROLLBACK_COMPLETE. An update with a corrected template is rejected. What should you do?
Question 30A baseline IAM role and AWS Config rules must exist in every account of the Workloads OU, including accounts created later, with no manual steps. What should you use?
Question 31An Auto Scaling group uses two /28 subnets. When it tries to scale from 18 to 30 instances, launches fail with an error about insufficient free IP addresses in the subnets. What fixes this?
Question 32Security wants a hardened, patched golden AMI built and tested every month and distributed to three Regions and five accounts automatically. What should you use?
Question 33cdk deploy fails in a newly created account because the CDK staging bucket and deployment roles do not exist. What should you do?
Question 34A stack creates an EC2 instance with a CreationPolicy. After 15 minutes it fails with "Received 0 SUCCESS signal(s) out of the 1 that was required". What should you check first?
Question 35You must run a shell script on 300 managed instances tagged Env=prod, no more than 10% at a time, and stop if five instances fail. SSH is blocked. What should you use?
Question 36A network team owns a central VPC. Application teams in other accounts of the same organization must launch their instances directly into its private subnets. What should the network team do?
Question 37An analyst's IAM policy allows s3:* on a bucket, but every request from her laptop returns AccessDenied. The bucket policy denies all access unless aws:SourceVpce equals the company VPC endpoint. Why?
Question 38The security team must list every S3 bucket, IAM role and KMS key that grants access to principals outside the company's AWS organization. What is the most efficient tool?
Question 39Auditors require that every S3 bucket in all accounts blocks public access, with violations fixed automatically and compliance shown in one place. What should you implement?
Question 40An auditor finds that the root EBS volume of a production instance is unencrypted. Policy requires encryption with a customer managed KMS key. How do you fix it?
Question 41You add an alternate domain name to a CloudFront distribution. The ACM certificate you issued in eu-west-1 for that name does not appear in the certificate list. What should you do?
Question 42An application reads its database password from a plain-text file on EC2. The password must be stored securely and rotated every 30 days without code redeployments. What should you use?
Question 43Data residency rules allow only eu-west-1 and eu-central-1. Any attempt to create resources elsewhere, in any member account, must be blocked. What should you use?
Question 44Instances in private subnets download several terabytes per month from S3 in the same Region through a NAT gateway. NAT gateway data processing charges are high. What reduces the cost?
Question 45VPC A is peered with VPC B, and VPC B is peered with VPC C. Routes and security groups look correct, yet instances in A cannot reach instances in C. Why?
Question 46On-premises servers connected by Direct Connect must resolve records in a Route 53 private hosted zone associated with a VPC. What should you set up?
Question 47For one week, the release team wants about 10% of users sent to a new stack behind a second load balancer and 90% kept on the current stack, with rollback by changing one value. Which Route 53 routing policy should be used?
Question 48A CloudFront distribution's cache hit ratio is below 10%. Its cache policy includes all headers, all cookies and all query strings in the cache key, although the origin returns the same content for most requests. What should you do?
Question 49A multiplayer game uses UDP servers in two Regions. Partners must allowlist fixed IP addresses, and players must move to the healthy Region within seconds of a failure. What should you use?
Question 50An EC2 instance cannot connect to a database instance on port 5432 in the same VPC. You must identify the blocking component from the configuration without generating traffic. What should you use?