Final Assessment :DevOps and AI on AWS: CI/CD for Generative AI Applications (DevOps and AI on AWS Specialization) Answers 2025
1. Question 1
How does CloudTrail support monitoring/troubleshooting?
-
❌ Runs commands on EC2
-
✅ Logs API activity + records resource changes
-
❌ Monitors performance metrics
-
❌ Automates patching
Explanation:
CloudTrail logs who did what, when, and from where — essential for audits.
2. Question 2 — (Select TWO)
Tools to troubleshoot AWS deployment issues
-
❌ AWS X-Ray (application tracing, not deployment debugging)
-
❌ S3 Buckets
-
✅ AWS CodeDeploy Deployment Logs
-
✅ AWS CloudFormation Change Sets
-
❌ Patch Manager
Correct Answers: 3 and 4
Explanation:
-
Deployment Logs → show why deployment failed
-
Change Sets → preview infra changes before execution
3. Question 3
Service best for detecting & logging config changes
-
❌ CloudTrail (API calls, not config states)
-
✅ AWS Config
-
❌ Systems Manager
-
❌ CloudWatch
Explanation:
AWS Config tracks configuration drift, compliance, and changes to resources.
4. Question 4 — (Select TWO)
Essential DevOps practices for deploying generative AI models
-
❌ Single environment, no rollback
-
❌ Replace IaC with manual config
-
✅ Automate model testing across diverse prompts
-
❌ Disable CD
-
✅ Monitor inference performance & resource usage
Correct Answers: 3 and 5
5. Question 5
Purpose of CodeDeploy lifecycle event hooks
-
✅ Automate deployment actions at specific stages
-
❌ Store env variables
-
❌ Define infrastructure
-
❌ Monitor EC2
Explanation:
Hooks define steps like BeforeInstall, AfterInstall, ApplicationStart, etc.
6. Question 6 — (Select TWO)
Deployment strategies that minimize release risk
-
❌ In-place
-
✅ Blue/Green Deployment
-
❌ All-at-once
-
✅ Canary Deployment
-
❌ Manual Deployment
Correct Answers: 2 and 4
7. Question 7
Why is prompt variability important?
-
✅ Small changes produce different outputs
-
❌ Responses are deterministic
-
❌ Tested only after infra
-
❌ Prompts produce identical results
8. Question 8
Key challenge integrating generative AI in DevOps
-
✅ Testing must handle variability + non-deterministic outputs
-
❌ AI does not require monitoring
-
❌ AI outputs deterministic
-
❌ CI incompatible with AI
9. Question 9
Major benefit of reviewing CloudFormation templates before execution
-
❌ Monitor resource use
-
❌ Test applications
-
✅ Ensure infra changes match requirements
-
❌ Speed up deployments
10. Question 10
Why are X-Ray service maps useful?
-
✅ Visualize flow of requests across components
-
❌ Execute commands
-
❌ Log API calls
-
❌ Monitor CPU/memory
🧾 Summary Table
| Q# | Correct Answer | Key Concept |
|---|---|---|
| 1 | Logs API activity | CloudTrail purpose |
| 2 | Deployment Logs, Change Sets | Troubleshooting |
| 3 | AWS Config | Config change detection |
| 4 | Auto prompt testing + monitor inference | GenAI DevOps |
| 5 | Automate stage actions | CodeDeploy hooks |
| 6 | Blue/Green, Canary | Safe deployments |
| 7 | Small changes → diff outputs | Prompt variability |
| 8 | Non-deterministic testing | GenAI CI/CD |
| 9 | Validate infra changes | CloudFormation |
| 10 | Visualize request flow | X-Ray maps |