AWS DevOps High Availability 5 — Questions and Answers
Question 1: A DevOps team uses AWS CodeDeploy with an Auto Scaling group. During a deployment, CodeDeploy fails on 3 out of 10 instances. What is the default behavior?
- Deployment continues on remaining instances and only failed instances are retried
- CodeDeploy stops the deployment and rolls back all instances based on the configured failure threshold (Correct answer)
- The Auto Scaling group terminates failed instances and launches replacements
- CodeDeploy marks the deployment successful if more than 50% of instances succeed
Correct answer: CodeDeploy stops the deployment and rolls back all instances based on the configured failure threshold
CodeDeploy evaluates the failure threshold and if exceeded, triggers an automatic rollback by redeploying the previous revision to all affected instances.
Question 2: Which S3 feature helps maintain availability of objects during multi-part upload failures or incomplete uploads that consume storage?
- S3 Object Lock with governance mode
- S3 Lifecycle policy to abort incomplete multipart uploads after a specified number of days (Correct answer)
- S3 Replication rules with delete marker replication
- S3 Requester Pays configuration
Correct answer: S3 Lifecycle policy to abort incomplete multipart uploads after a specified number of days
A lifecycle rule targeting incomplete multipart uploads automatically cleans them up after a configured number of days, freeing storage and reducing confusion.
Question 3: A high-availability architecture requires Aurora MySQL instead of standard RDS MySQL. What is a key availability advantage Aurora provides?
- Aurora stores data in a single AZ with daily snapshots
- Aurora automatically replicates storage across 3 AZs with 6 copies of data and supports up to 15 read replicas with sub-10ms failover (Correct answer)
- Aurora requires manual promotion of read replicas during failover
- Aurora uses magnetic storage for higher durability
Correct answer: Aurora automatically replicates storage across 3 AZs with 6 copies of data and supports up to 15 read replicas with sub-10ms failover
Aurora's distributed storage layer maintains 6 copies across 3 AZs and its failover to a read replica typically completes in under 30 seconds, often under 10 seconds.
Question 4: In an AWS Step Functions state machine used for deployment orchestration, which feature helps handle transient failures in downstream service calls?
- Activity tasks with heartbeat timeouts
- Built-in retry and catch configurations on task states with exponential backoff (Correct answer)
- Express workflows with synchronous execution
- State machine versioning with alias routing
Correct answer: Built-in retry and catch configurations on task states with exponential backoff
Step Functions allows defining Retry policies with maxAttempts and backoffRate directly on task states, handling transient errors without custom retry logic.
Question 5: An engineering team needs to test their application's resiliency to EC2 instance failures. Which AWS service enables controlled fault injection experiments?
- AWS Config with remediation actions
- AWS Fault Injection Simulator (FIS) with EC2 instance stop/terminate actions (Correct answer)
- AWS Systems Manager Run Command with stop-instance document
- CloudWatch Synthetics canaries with failure simulation
Correct answer: AWS Fault Injection Simulator (FIS) with EC2 instance stop/terminate actions
AWS FIS provides controlled chaos engineering experiments including EC2 instance termination, CPU stress, network disruption, and other fault injections with safety guardrails.
Question 6: A company wants to ensure its Elastic Load Balancer distributes traffic evenly across EC2 instances in two AZs even when one AZ has fewer instances. Which feature enables this?
- ALB slow start mode
- Cross-zone load balancing (Correct answer)
- Least outstanding requests routing algorithm
- ALB weighted target groups
Correct answer: Cross-zone load balancing
Cross-zone load balancing distributes requests evenly across all registered targets regardless of which AZ they are in, preventing AZ-level traffic imbalance.
Question 7: When using AWS CodePipeline with a multi-region deployment, which service must be configured in each target region to store pipeline artifacts?
- Amazon EFS with cross-region replication
- An Amazon S3 artifact bucket in each target region with appropriate KMS key permissions (Correct answer)
- AWS CodeCommit repository mirrored to each region
- Amazon ECR repository replicated across regions
Correct answer: An Amazon S3 artifact bucket in each target region with appropriate KMS key permissions
CodePipeline requires a regional S3 artifact bucket and matching KMS key in every region where cross-region actions are configured to store and encrypt pipeline artifacts.
AWS DevOps High Availability 5
A DevOps team uses AWS CodeDeploy with an Auto Scaling group.
During a deployment, CodeDeploy fails on 3 out of 10 instances.
What is the default behavior?