For years this exam rewarded knowing the console. It still does, and the current guide now names Terraform, Config Connector, Gemini CLI and Gemini Cloud Assist among the tools you plan and build with. Google publishes no passing score either. Results arrive as pass or fail with nothing to say which section let you down, which is why Associate Cloud Engineer exam questions are worth working one section at a time. Where a date is already in the calendar, run whole sets inside two hours so pace stops being a variable. Where the date is still open, take the four sections against what you have actually built.
| Exam name | Google Associate Cloud Engineer |
| Exam code | ASSOCIATE-CLOUD-ENGINEER |
| Certification | Google Cloud Certified |
| Time allowed | 2 hours |
| Passing score | Not published. Google reports pass or fail only. |
| Recommended experience | 6+ months hands-on with Google Cloud. No prerequisites. |
| Exam cost | USD 125, plus tax where applicable |
| Languages | English, Japanese, Spanish, Portuguese |
| Retake policy | 14 days after a first failure, 60 days after a second, 365 days after a third. Four attempts in any two years, paid for each time. |
| Certification validity | Three years. Renew by sitting the exam again, by taking the one-hour renewal exam at USD 75, or through designated Google Skills courses. |
| Practice questions in our bank | 363 |
| Questions on the real exam | 50 |
A sample from the bank, written to the four sections of Google's current exam guide.
Your company plans to migrate its on-premises PostgreSQL database to Google Cloud. The workloads are demanding, requiring fast transactional and analytical performance. You need to select a fully managed database service on Google Cloud. Your solution must also be able to synchronously replicate and optimize the storage layer. What should you do?
Correct Answer: B
Comprehensive and Detailed Explanation From Exact Extract:
The requirements point to the highest performance tier for a managed PostgreSQL-compatible database, specifically citing demands for transactional and analytical performance, synchronous replication, and storage optimization.
AlloyDB for PostgreSQL is Google Cloud's highly specialized, fully managed database that is optimized for demanding enterprise workloads. It is superior to standard Cloud SQL for performance and architectural features like synchronous replication and an optimized storage layer, directly matching all the stated requirements.
Database Migration Service (DMS) is the recommended tool for performing the migration itself.
'AlloyDB for PostgreSQL is a fully managed PostgreSQL-compatible database service that is designed for demanding enterprise workloads... [It] delivers high availability and synchronous replication for better data protection and faster failover.'
'Database Migration Service (DMS) simplifies your database migrations to Cloud SQL and AlloyDB.'
(You are migrating your company's on-premises compute resources to Google Cloud. You need to deploy batch processing jobs that run every night. The jobs require significant CPU and memory for several hours but can tolerate interruptions. You must ensure that the deployment is cost-effective. What should you do?)
Correct Answer: D
Spot VMs (formerly known as preemptible VMs) are Compute Engine virtual machine instances that are available at a much lower price than standard Compute Engine instances. However, Compute Engine might preempt (stop) these instances if it needs to reclaim those resources for other tasks. This makes Spot VMs ideal for batch processing jobs that are fault-tolerant and can handle interruptions, as they can be restarted when resources become available again. This directly addresses the requirement for a cost-effective solution for interruptible workloads.
Option A: While containerization offers portability and consistency, it doesn't inherently provide cost savings for compute resources. You would still need to choose a cost-effective underlying compute option.
Option B: Custom machine types allow you to tailor the CPU and memory configuration of your VMs, which can optimize costs to some extent by avoiding over-provisioning. However, they don't offer the significant cost reduction that Spot VMs provide.
Option C: The M1 machine series is a specific family of Compute Engine instances optimized for memory-intensive workloads. While potentially suitable for the job's requirements, it doesn't inherently address the cost-effectiveness requirement as directly as Spot VMs, which are priced lower regardless of the machine series.
Reference to Google Cloud Certified - Associate Cloud Engineer Documents:
The concept and use cases for Spot VMs are explicitly covered in the Compute Engine section of the Google Cloud documentation, which is a key area for the Associate Cloud Engineer certification. The cost savings and suitability for fault-tolerant workloads are highlighted as primary benefits.
Your team is building a website that handles votes from a large user population. The incoming votes will arrive at various rates. You want to optimize the storage and processing of the votes. What should you do?
Correct Answer: B
Pub/Sub is a scalable and reliable messaging service that can handle large volumes of data from different sources at different rates. It allows you to decouple the producers and consumers of the data, and provides a durable and persistent storage for the messages until they are delivered. Cloud Functions is a serverless platform that can execute code in response to events, such as messages published to a Pub/Sub topic. It can scale automatically based on the load, and you only pay for the resources you use. By using Pub/Sub and Cloud Functions, you can optimize the storage and processing of the votes, as you can handle the variable rates of incoming votes, process them in real time or near real time, and avoid managing servers or VMs.Reference:
Pub/Sub documentation
Cloud Functions documentation
Choosing a messaging service for Google Cloud
Your development team needs a new Jenkins server for their project. You need to deploy the server using the fewest steps possible. What should you do?
Correct Answer: D
Your company uses Pub/Sub for event-driven workloads. You have a subscription named email-updates attached to the new-orders topic. You need to fetch and acknowledge waiting messages from this subscription. What should you do?
Correct Answer: C
The goal is to pull (fetch) messages from a subscription and acknowledge them.
The gcloud pubsub subscriptions **pull** command retrieves messages from a specified subscription.
The --auto-ack flag instructs the command to automatically acknowledge the messages after they are successfully retrieved, combining the two required actions into one command.
'Pulls one or more messages from the specified subscription. To acknowledge the pulled messages, use the --auto-ack flag.'
The four sections below come from Google's own exam guide, weighted as Google weights them, each with a note on what it actually asks of you.
Setting up a cloud solution environment
~20%Most of this happens before anything is deployed. You are asked where a project sits in a resource hierarchy, which organization policy reaches it, and who gets an IAM role at which level. Billing arrives as configuration rather than finance, so budgets, alerts, exports, and which account a project is linked to.
Planning and implementing a cloud solution
~30%The section that keeps asking you to choose. A workload arrives with a shape and you pick the compute, the storage and the data product that suit it. Networking sits here as well, from VPCs and firewall policies through to load balancers, alongside the tooling you build with, Terraform and Gemini CLI included.
Ensuring the successful operation of a cloud solution
~30%Everything here is done to something already running. Snapshots, node pools, traffic splits across Cloud Run revisions, lifecycle rules on a bucket, a database restored from backup. Monitoring and logging take a large share, and the questions want to know which tool shows what rather than how to read the graph it draws.
Configuring access and security
~20%Identity, and very little else. A fifth of the exam turns on IAM policies, where a role is inherited from, and what a custom role exists to do. Service accounts carry the rest of it, impersonation and short-lived credentials among them, along with how a GKE workload gets an identity without a key file on disk.
Checked 2026 — : The current guide runs to four sections, weighted 20, 30, 30 and 20. Gemini CLI, Gemini Cloud Assist, Google Antigravity, Cloud NGFW and Agent Runtime are named among the tools and services in scope.
Earlier guide — : Five sections at 20, 17.5, 25, 20 and 17.5, with planning and deploying counted separately. Google changed it on June 30, 2026. Google has removed the dates from their dynamic document.
Source: Google's Associate Cloud Engineer exam guide. Our question bank is updated to match each revision.
Every question is a choice between products that would all technically work, starting with Cloud Run against GKE, Persistent Disk against Filestore, and a custom IAM role against a predefined one. Google never asks what these are, only which one belongs in the situation described, and it asks that as multiple choice or multiple select. Recognising a service name gets you nowhere on a paper built entirely out of comparisons.
Where a permission comes from
Tested via: multiple-choice questionsConfiguring access and security is 20% of the exam, and Setting up carries the hierarchy the whole thing hangs off, another 20%. A role granted at the organization level reaches every project beneath it, and that inheritance is what most wrong answers overlook. Service accounts add a second layer, because a workload's identity and a person's identity are granted in different ways. Impersonation and short-lived credentials turn up as the safer alternative to a downloaded key.
Which product fits which shape of workload
Tested via: multiple-choice and multiple-select questionsCompute Engine, GKE, Cloud Run and Cloud Run functions all run code, and the question is which one the described workload belongs on. Planning and implementing is 30% of the exam, and that decision, made over and over, is most of it. Storage behaves the same way, with Cloud Storage classes, Persistent Disk, Filestore and Hyperdisk each having a case they win. Data adds Cloud SQL, Spanner, Bigtable, Firestore, AlloyDB and BigQuery, where the deciding detail is usually consistency, scale or query shape.
What you change on something already running
Tested via: multiple-select questionsA node pool gets resized, a Cloud Run revision takes a percentage of traffic, a bucket picks up a lifecycle rule, a Cloud SQL instance is restored to an earlier point. All of that is Ensuring the successful operation, joint-largest at 30%, and none of it starts from an empty project. Monitoring runs through the section, and the questions tend to ask which of Cloud Logging, Cloud Monitoring, Cloud Trace or the log router is the right place to look.
Where a network begins and ends
Tested via: multiple-choice questionsNetworking is planned in one section and managed in another, which is why it feels larger than any single weighting suggests. A VPC's subnets, its custom routes and its firewall rules decide what can reach what, and Cloud NGFW policies now carry that with tags and service accounts as attributes. Connectivity is the other half, where Cloud VPN, VPC Network Peering and Cloud Interconnect each answer a different question about what you are joining to.
Sitting soon is usually the plan, so here is what a failure costs on this one. Another USD 125, then 14 days before a second try, 60 days after a second failure and 365 after a third, with four attempts allowed in any two years.
and Why Prefer Pass4Success Practice Material
Failing this one costs USD 125 and a fortnight, and the wait stretches to 60 days if it happens twice and 365 if it happens three times. Against a ceiling like that, what practice material costs matters less than whether it follows the four-section guide Google is testing from now.
Follows the guide Google is testing from now
Google's guide moved from five sections to four and picked up Gemini tooling, Cloud NGFW and Agent Runtime along the way. Material written to the older structure still looks complete, and that is the trouble with it.
✓ Ours: the bank runs on its own update cycle, tracking how Google actually tests rather than waiting for a guide revision to force the issue. The date it was last updated sits in the table at the top of this page.
Questions you can read before you spend anything
Cloud certifications attract more recycled material than most, partly because the product names change often enough to make anything look current. Reading a few questions settles it quickly.
✓ Ours: both formats have a free demo, so nothing has to be taken on description. The samples higher up this page are drawn from the same bank.
Two ways to work, because two hours is a long sitting
Reading through a section at a desk and holding concentration across 60 questions are separate abilities, and this exam asks for the second one for two hours straight.
✓ Ours: the PDF and the practice test come together in one purchase, and the practice test runs browser-based or as a Windows desktop version.
Coverage across all four sections, not just the popular two
Compute and storage questions are the ones every free set has. Billing configuration, organization policy and log routing are where a thin bank runs out, and together those account for a good deal of the exam.
✓ Ours: 363 Associate Cloud Engineer questions spread across the guide's four sections.
A price that reads sensibly against USD 125
Google's fee is fixed and public, which makes it an easy figure to measure preparation against. A second sitting doubles it and adds a fortnight you had not planned for.
✓ Ours: USD 69 Practice material Compare that to a single USD 125 for retake plus 14 days wait before rebook you exam, and the math favors getting it right the first attempt.
Below, candidates set out which product comparisons came up, how the two hours felt in practice, and what they wish they had spent longer on. Add yours once you have sat it.
The Associate Cloud Engineer exam has 50 to 60 questions and a two-hour limit. That leaves a little over two minutes each, which is generous by certification standards and deliberately so, because most items describe a situation before they ask anything. Google publishes a range rather than a fixed number, so the length varies between candidates.
Google does not publish one. The exams are built to decide only whether a candidate meets a minimum standard, so results come back as pass or fail with no score and no section breakdown. A provisional result appears on screen the moment you submit, and Google confirms it within 7 to 10 days.
Registration is USD 125, plus tax where it applies. Every attempt is paid for separately, so a failure means paying again rather than rebooking for nothing. The renewal exam later costs USD 75 for one hour and 20 questions, which is worth knowing when you are budgeting three years ahead rather than three weeks.
No code exists for it. Google does not assign alphanumeric exam codes the way CompTIA and Microsoft do, so there is nothing to search for beyond the certification name itself. You register as Associate Cloud Engineer through Google's own registration system. Anything advertising a code for this exam has invented one.
Associate Cloud Engineer is the usual first step for anyone who works with Google Cloud directly. Cloud Digital Leader sits below it and is written for non-technical roles. Above it are the Professional certifications, Cloud Architect and Cloud Developer among them, which stay valid for two years rather than three and assume the ground this exam covers.
There are no prerequisites, and Google recommends six months or more of hands-on work with Google Cloud. That recommendation matters more than it sounds. The questions are written as choices between real products rather than as definitions, and product choice is the thing that reads differently once you have deployed something yourself.
The Google Cloud Certified - Associate Cloud Engineer exam uses multiple choice and multiple select questions only. There are no labs, no simulated consoles and no drag and drop. What makes it harder than that sounds is that several options are usually workable, and the exam wants the one Google would choose for the situation described.
Studying products in isolation is the expensive one, because nearly every question is a comparison rather than a recall. After that comes underestimating IAM inheritance, where a role granted high in the hierarchy quietly reaches everything beneath it. The third is skipping billing and monitoring, which feel like administration and carry real weight.
The certification lasts three years from the date it is awarded. Renewal opens 180 days before it expires and there are three routes to it. You can sit the full exam again, take the one-hour renewal exam of 20 questions at USD 75, or complete designated courses and skill badges through Google Skills.
Yes, after a wait that grows with each attempt. Google requires 14 days before a second attempt, 60 days before a third and 365 days before a fourth, and caps candidates at four attempts in any two-year window. Each one is paid for in full, so a casual first attempt costs more than it appears to.
Full access to all 363 Associate Cloud Engineer questions, in both the PDF and the practice test.
Get Premium AccessThis page is checked against Google's published exam guide, which currently runs to four sections and carries no version number or date of its own. The bank keeps its own schedule rather than waiting for a guide revision, because Google renames and replaces products far more often than it restructures an exam. When the guide does change, affected Associate Cloud Engineer questions are re-checked and the update date above moves.