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. Engineering tagged every resource with CostCenter a week ago. Cost Explorer still can't group
spend by that tag. What is the most likely cause?
2. A sandbox account must be prevented from launching new resources as soon as its monthly spend passes $2,000, without anyone watching. What do you configure?
3. Finance needs hourly, per-resource cost data loaded into Athena and joined to an internal asset database. Which source?
4. A job writes 5 KB status files to S3 and deletes them after 10 days. To cut cost, a developer moves them to S3 Standard-IA. What happens to the bill?
5. Medical images are rarely viewed — about once a quarter — but must display in milliseconds when a clinician opens them. Lowest-cost class?
6. A bucket's access pattern is unknown and will change over time. Objects are mostly 2–50 MB. Which class needs no lifecycle design and charges no retrieval fees?
7. A research institute publishes a 400 TB dataset in S3 for other AWS customers. It wants those customers, not itself, to pay for their downloads. What does it enable, and what does it give up?
aws:SourceAccount; it gives up versioning8. A private-subnet gp2 volume is 2 TB but holds 200 GB of data; it was sized up only to get IOPS.
What is the cost-optimal change?
sc1gp3 at a smaller size with the IOPS provisioned separatelyst1 to get more throughput per dollar9. A company runs a steady fleet on EC2 today, plans to move half of it to ECS on Fargate next year, and may change instance families. It wants the largest discount that survives those changes. What does it buy?
10. Which statement about Spot Instances is correct?
11. A workload needs a discount and guaranteed capacity in us-east-1a for three years. Which
purchase meets both requirements?
12. A company is moving Windows Server and SQL Server workloads that use existing per-core licences. It needs visibility of sockets and cores. What do you choose?
13. Application servers take 25 minutes to warm their in-memory caches. The team wants to stop them overnight to save money without losing that warm state. What do you use?
14. A new DynamoDB table's traffic is unknown and expected to be spiky. Which capacity mode does AWS describe as the default and recommended option?
15. An application does 100 strongly consistent reads per second of 6 KB items. How many RCUs does provisioned mode need?
16. A regulated team must keep monthly RDS backups for seven years at the lowest cost. What do you recommend?
17. Instances in private subnets pull several TB a month from S3 in the same Region through a NAT gateway. What cuts that cost most directly?
18. A development VPC spans three AZs with low outbound traffic. A production VPC spans three AZs, must survive an AZ failure, and has heavy outbound traffic. Which NAT design fits each?
19. Two VPCs in the same Region exchange tens of TB a month. There are no other VPCs and no plans for more. Lowest-cost connection?
20. An on-premises site needs more than 1.25 Gbps to AWS over Site-to-Site VPN while a Direct Connect circuit is being ordered. What design delivers it?
1 — B. "You must activate both types of tags separately before they can appear in Cost Explorer", and "Only the management account in an organization … ha[s] access to the cost allocation tags manager." A is false — user-defined tags are supported. D is false — tags "can take up to 24 hours to appear", not a month.
2 — C. A budget action can "run either automatically or after your manual approval", and actions include applying "a service control policy (SCP)". B only notifies. A and D don't act on anything.
3 — C. CUR "contains the most comprehensive set of cost and usage data available", publishes to "an Amazon S3 bucket that you own", can be broken down "by the hour" and "by product or product resource", and integrates with Athena. Cost Explorer is interactive analysis, not a data feed.
4 — B. "If an object is less than 128 KB, Amazon S3 charges you for 128 KB. If you delete an object before the end of the 30-day minimum storage duration period, you are charged for 30 days." A 5 KB, 10-day object is billed as 128 KB for 30 days. D is also false — Standard-IA has per-GB retrieval fees.
5 — C. Glacier Instant Retrieval is for data "accessed once a quarter with millisecond access". A and B are archival — objects must be restored first. D is designed for monthly access and isn't resilient to AZ loss.
6 — B. Intelligent-Tiering is for "unknown, changing, or unpredictable access patterns", with "no retrieval fees". The objects are well above the 128 KB monitoring threshold, so they will tier.
7 — A. "The requester instead of the bucket owner pays the cost of the request and the data download" — and "If you enable Requester Pays on a general purpose bucket, anonymous access to that bucket is not allowed." B is wrong: "The bucket owner always pays the cost of storing data."
8 — B. gp3 offers "a 20 percent lower price per GiB" and lets you "scale volume performance
independently of volume size", and the change is made with Elastic Volumes "without interrupting" the
instance. A and C are HDD types built for throughput, not IOPS.
9 — C. Compute Savings Plans apply "regardless of instance family … Region … operating system" and "also apply to your Fargate and Lambda usage". A is locked to "a specific instance family in a chosen AWS Region". B can't follow the workload to Fargate. D is wrong on the facts — "Spot Instances are not covered by Savings Plans."
10 — B. The notice is "issued two minutes before Amazon EC2 stops or terminates your Spot Instance", but with hibernation "you do not receive a two-minute warning because the hibernation process begins immediately." A: Spot spend "does not apply the commitments". C: notices are "best effort". D: the price is "adjusted gradually based on the long-term supply of and demand".
11 — C. "A zonal Reserved Instance reserves capacity in the specified Availability Zone." A regional RI "does not reserve capacity", and Savings Plans are a pricing commitment only.
12 — B. Dedicated Hosts support "per-socket, per-core, or per-VM software licenses" and provide "visibility of the number of sockets and physical cores". Dedicated Instances have "No visibility" and only "Partial support" for BYOL.
13 — B. Hibernation "saves the contents from the instance memory (RAM)" and AWS names the use case: an instance that "takes a long time to bootstrap and build a memory footprint". You're "not charged for instance usage" while hibernated and stopped. D loses the warm state.
14 — C. "On-demand mode is the default and recommended throughput option for most DynamoDB workloads." Provisioned suits "steady workloads with predictable growth".
15 — C. One RCU covers one strongly consistent read per second "for an item up to 4 KB", rounded up per item: 6 KB → 2 RCU per read × 100 = 200. With eventually consistent reads it would be 100 (A).
16 — B. Automated backup retention is "between 0 and 35 days", so A is impossible. AWS Backup supports long retention and cold storage (with a 90-day minimum in cold). C and D are availability features, not retention.
17 — B. Gateway endpoints connect to S3 "without requiring an internet gateway or a NAT device", and "There is no additional charge for using gateway endpoints." NAT gateways charge per GB processed; this removes that traffic from the NAT. A trades per-GB charges for operational overhead. C adds cost.
18 — B. For dev, one NAT gateway minimises hourly charges and the AZ-loss risk is acceptable. For production, AWS says "Create a NAT gateway in each Availability Zone to ensure zone-independent architecture", and per-AZ gateways avoid cross-AZ transfer for heavy traffic.
19 — B. "VPC peering offers the lowest overall cost and highest aggregate performance when compared to other options for inter-VPC connectivity", with no per-GB processing fee. Transit Gateway adds per-attachment hours and $0.02/GB processed (US East (Ohio), 2026-09-25).
20 — B. "For Site-to-Site VPN connections on a transit gateway, you can use ECMP to get higher VPN bandwidth by aggregating multiple VPN tunnels. To use ECMP, the VPN connection must be configured for dynamic routing." Each standard tunnel is "Up to 1.25 Gbps".
| Score | Read this as |
|---|---|
| 18–20 | Domain 4 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 conditions attached to the cheap options. Redo lessons 2 and 3. |
| < 10 | Work the module in order and do the lab before re-testing. |
Miss patterns: