Pair your devices with a code and playback position follows you: pause on this device, hit resume on the other. Position is saved to the site every minute and on pause.
Open this panel on your other device and enter the same code.
20 questions, four options each, one correct answer, no partial credit.
⚠️ Exam-styled questions written from the task statements and AWS documentation. Not the AWS Official Practice Question Set and not calibrated against real exam items.
1. Three downstream systems must each process every order event. None may be lost if a system is down for an hour. What architecture?
2. A workflow must pause for a manager's approval before continuing. Which Step Functions workflow type, and why?
3. A Step Functions workflow processes 60,000 IoT events per second, each in under a second. Which type?
4. Which statement about scaling is correct?
5. An Auto Scaling group has min 2, max 6, desired 6. CPU is pinned at 98% and no new instances launch. What is the most likely explanation?
6. What does an Auto Scaling group with min = max = desired = 3 provide?
7. Long-running batch jobs are being terminated when the Auto Scaling group scales in. What is the correct fix?
8. A requirement is to route /api/* and /static/* to different target groups. Which load
balancer?
9. A partner's firewall team must allowlist two fixed IP addresses for your endpoint. What do you provide?
10. AZ-A has 2 registered targets; AZ-B has 8. The 2 targets in AZ-A are overloaded. Why?
11. Which is a category error in a set of exam options?
12. An ECS service on Fargate is registered behind an ALB but the target group registers nothing. What is wrong?
instance instead of ip13. A nightly ETL job runs for about 40 minutes. Which compute option is ruled out, and why?
14. What is the durable reason to choose EKS over ECS?
15. Read queries are overwhelming an RDS instance, which is also a single point of failure. What do you implement?
16. Which Route 53 configuration produces active-passive failover?
17. You are creating Route 53 records pointing at an Application Load Balancer. What should you do about health checking?
18. A Lambda-backed API fails under load with database connection errors. What is the targeted fix?
19. The DR Region has data replicating continuously and core infrastructure provisioned, but the application servers are switched off. Which strategy is this, and what is the test that identifies it?
20. Which failover mechanism does AWS describe as a control plane operation and therefore less resilient?
1 — C. SNS gives every subscriber a copy; the SQS queue in front of each consumer holds the message while that consumer is down. A delivers each message to only one consumer. B and D attempt delivery and eventually give up — nothing stores the message for an hour-long outage.
2 — B. "Express Workflows only support Request Response integrations." Human approval uses the Wait for Callback pattern with a task token, which is Standard-only. D is also wrong on the facts: Express maxes out at five minutes.
3 — B. Express is "ideal for high-event-rate workloads, such as streaming data processing and IoT data ingestion" at "100,000 per second execution rate", against Standard's 2,000. A is wrong because Standard's rate ceiling can't carry 60,000/s.
4 — B. Horizontal scaling means any request can land on any instance, so state cannot live in instance memory. A is the classic error — vertical scaling leaves you with one instance. C describes vertical scaling. D is nonsense.
5 — B. "Amazon EC2 Auto Scaling ensures that your group never goes above this size." Desired has
already reached max. Raise max.
6 — B. It "replaces terminated or impaired instances to maintain your desired capacity" and "balances your instances evenly across the Availability Zones". Self-healing without scaling is a legitimate, common design. D is wrong — an ASG does not span Regions.
7 — C. AWS names both: "scale-in protection or custom termination policies to prevent instances with long-running processes from terminating early." A throws away elasticity to fix a configuration problem.
8 — B. "Path conditions. You can configure rules for your listener that forward requests based on the URL in the request." An NLB works at layer 4, where there is no URL to inspect.
9 — B. "you can optionally associate one Elastic IP address per subnet." An ALB's addresses change over time, which is precisely why you reach it through an alias record rather than a fixed IP.
10 — B. "By default, each load balancer node distributes traffic across the registered targets in its Availability Zone only." Traffic splits ~50/50 between the two nodes, so 2 targets share what 8 targets share. Fix with cross-zone load balancing or balanced target counts.
11 — B. ECS is the orchestrator; Fargate is the capacity it runs on. They are on different axes. A, C and D are all genuine comparisons.
12 — A. "you must choose ip as the target type, not instance. This is because tasks that use the
awsvpc network mode are associated with an elastic network interface, not an Amazon EC2 instance."
13 — C. Lambda is "Up to 15 minutes per invocation". A misattributes that limit to Fargate, which has no such ceiling.
14 — B. "Amazon EKS is certified Kubernetes-conformant, so you can deploy Kubernetes-compatible applications without refactoring and use Kubernetes community tooling and plugins." A is backwards — EKS has per-cluster pricing. C is false. D is false — ECS Anywhere exists.
15 — C. Two problems, two features. "A Multi-AZ standby cannot serve read requests. Multi-AZ deployments are designed to provide enhanced database availability and durability, rather than read scaling benefits." Read replicas are asynchronous with manual promotion, so they don't cover availability either.
16 — C. "you configure active-passive failover using the failover routing policy", and "active-active failover using any routing policy … other than failover." A, B and D all give active-active.
17 — B. "If you're routing traffic to any AWS resources that you can create alias records for, don't create health checks for those resources. When you create the alias records, you set Evaluate Target Health to Yes instead."
18 — C. RDS Proxy "establishes a database connection pool and reuses connections" and "queues or throttles application connections that can't be served immediately". Lambda scales by adding execution environments, each opening its own connection — a bigger instance (A) just moves the ceiling, and B and D address different problems.
19 — B. "pilot light cannot process requests without additional action taken first, whereas warm standby can handle traffic (at reduced capacity levels) immediately." That question — can it serve a request right now? — is the whole distinction.
20 — C. "be aware this is a control plane operation and therefore not as resilient as the data plane approach using Amazon Application Recovery Controller." A and B are both explicitly described as data plane operations; D requires no operation at all.
| Score | Read this as |
|---|---|
| 18–20 | Domain 2 is solid. Move on; revisit a week before the exam. |
| 14–17 | Good. Re-read the lesson behind each miss — the explanations name them. |
| 10–13 | You know the services but not the trade-offs. Redo lessons 2 and 6, then the lab. |
| < 10 | Work the module in order and do the lab before re-testing. |
Miss patterns: