Customize this training

Get a certificate — $16.67$5

Google Cloud Associate Cloud Engineer Exam Prep

Google Cloud Associate Cloud Engineer Exam Prep

Revise the four Associate Cloud Engineer sections: projects, billing and identity, compute, storage and networking, day-2 operations and monitoring, IAM and service accounts, with exam-style scenarios.

What you will learn

  • Describe the exam and set up a resource hierarchy, organisation policies, Cloud Identity users and groups, APIs and quotas
  • Manage billing accounts, budgets, alerts and exports, and prepare gcloud, observability and regional availability for a new environment
  • Choose and deploy the right compute option: Compute Engine, managed instance groups, GKE, Cloud Run and Cloud Run functions
  • Select storage, database and network designs, including VPCs, firewall policies, hybrid connectivity, load balancers and Infrastructure as Code
  • Operate running compute: snapshots, images, GKE node pools and autoscaling, Cloud Run revisions and traffic splitting
  • Operate storage, data and networks, and use Cloud Monitoring and Cloud Logging to alert, route logs and diagnose problems
  • Configure IAM roles, policy inheritance, custom roles and service accounts with least privilege and without long-lived keys

Modules

  1. The exam, the resource hierarchy and identities
  2. Billing, tooling and a ready-to-use environment
  3. Planning and deploying compute
  4. Planning storage, data and networking
  5. Operating compute: VMs, GKE and Cloud Run
  6. Operating data and networks, monitoring and logging
  7. IAM and service accounts

Course facts

Level
Exam preparation
Estimated study time
about 2 hours
Final assessment
44 questions · Pass mark 70%
Available in
English, Français
Price
Free. Optional verifiable certificate: $5
Last updated
Published by
AMAADOR ACADEMY

Who this course is for

Cloud and system administrators, junior cloud engineers, developers moving into operations and IT staff from other clouds who are preparing for the Google Cloud Associate Cloud Engineer exam.

Prerequisites

Basic Linux command line, TCP/IP and DNS, and some familiarity with containers. Google recommends 6+ months of hands-on Google Cloud experience: practise each topic in a free-tier or sandbox project with a billing budget set.

Course content

1. The exam, the resource hierarchy and identities

The facts below come from Google Cloud's own certification page and the current Associate Cloud Engineer exam guide (checked 18 November 2026). Check them again before you book: Google revises the guide and renames products often.

ItemStandard exam
Length2 hours
Questions50-60 multiple choice and multiple select
FeeUS$125 plus tax where applicable
LanguagesEnglish, Japanese, Spanish, Portuguese
DeliveryOnline-proctored or at a test centre
Validity3 years; a shorter 1-hour, 20-question renewal exam (US$75) is offered from 180 days before expiry
Recommended6+ months of hands-on Google Cloud experience; no formal prerequisite

Google does not publish a pass mark or a score scale for this exam. This course uses 70% as its own practice target.

The four sections and their weights

  1. Setting up a cloud solution environment — about 20% (modules 1-2)
  2. Planning and implementing a cloud solution — about 30% (modules 3-4)
  3. Ensuring successful operation of a cloud solution — about 30% (modules 5-6)
  4. Configuring access and security — about 20% (module 7)

The current guide adds AI tooling throughout: Gemini Cloud Assist, Gemini CLI, Application Design Center, and Agent Runtime on Gemini Enterprise Agent Platform (formerly Vertex AI Agent Engine). Expect product names in questions to match the current console.

2. Billing, tooling and a ready-to-use environment

A Cloud Billing account pays for one or more projects. A project is linked to at most one billing account at a time; without a link, paid services stop. Organisations often have one billing account per legal entity or cost centre, or one central account for everything.

RoleWhat it allows
Billing Account AdministratorManage the billing account, payment and users on it
Billing Account UserLink projects to this billing account (combined with project rights)
Billing Account ViewerSee costs and transactions, change nothing
Project Billing ManagerLink or unlink a given project to a billing account

To link a project you need permission on both sides: the project (for example Project Billing Manager or Owner) and the billing account (Billing Account User or Administrator).

Budgets and alerts

  1. Scope — the whole billing account, or selected projects, services or labels
  2. Amount — a fixed figure or last month's spend
  3. Thresholds — for example 50%, 90% and 100%, on actual or forecasted spend
  4. Notify — email billing admins, Cloud Monitoring channels, or publish to a Pub/Sub topic for automation

3. Planning and deploying compute

Most compute questions are a choice between five platforms. Start from what the workload needs and how much you want to manage.

OptionYou manageTypical signal in the question
Compute EngineOS, patches, scaling setupLift and shift, specific OS or licence, custom kernel, GPUs on a VM
GKE StandardNodes, node pools, clusterKubernetes, need for node-level control or DaemonSets
GKE AutopilotWorkloads only; Google runs the nodesKubernetes with the least operations work, pay for Pod requests
Cloud RunThe container imageStateless HTTP container, scale to zero, pay per use
Cloud Run functionsA single functionSmall piece of code reacting to an event (Pub/Sub, Storage, HTTP)

Remember: when the question says "least management overhead" or "fully managed" and the app is a stateless container, Cloud Run is usually right. When it says "Kubernetes" and "minimal node management", think Autopilot.

GPUs or TPUs?

  • GPUs attach to Compute Engine VMs or GKE nodes: general accelerated work such as ML training and inference, rendering, video transcoding, CUDA code.
  • TPUs are Google-designed accelerators for large-scale ML training and inference with frameworks such as JAX, TensorFlow or PyTorch/XLA.
  • Accelerators are available only in some zones and need their own quota.

4. Planning storage, data and networking

NeedProduct
Relational, MySQL / PostgreSQL / SQL Server, one regionCloud SQL (enable high availability for a standby in another zone)
PostgreSQL-compatible with high performance for mixed transactional and analytical workAlloyDB for PostgreSQL
Relational, horizontally scalable, strongly consistent, multi-regionSpanner
Huge volume of key-based reads and writes, low latency (IoT, time series)Bigtable
Document database for web and mobile appsFirestore
Serverless analytics with SQL over large datasetsBigQuery
In-memory cacheMemorystore
Asynchronous messaging, decouplingPub/Sub (or Managed Service for Apache Kafka when Kafka is required)
Stream or batch processing pipelines (Apache Beam)Dataflow

Files and objects

  • Cloud Storage: objects in buckets. Location: region, dual-region or multi-region.
  • Filestore: managed NFS file shares for VMs and GKE. NetApp Volumes: managed NFS/SMB with NetApp features. Managed Lustre: parallel file system for HPC and AI training.
Storage classMinimum storage durationTypical access
StandardnoneHot data, frequent access
Nearline30 daysAbout once a month
Coldline90 daysAbout once a quarter
Archive365 daysLess than once a year, compliance copies

5. Operating compute: VMs, GKE and Cloud Run

Day-2 tasks are what you do after launch: connect, inspect, protect and change running machines.

TaskHow
List running VMsgcloud compute instances list (add --filter=status=RUNNING)
Connect by SSHgcloud compute ssh VM --zone ZONE (add --tunnel-through-iap if there is no external IP)
Connect to WindowsSet a Windows password, then RDP (through IAP if private)
Attach a GPUStop the VM, change the machine type or add the accelerator, set maintenance to terminate, install drivers

Snapshots, images and machine images

  • Snapshot: a point-in-time backup of one disk. Snapshots are incremental and can be stored multi-regionally. Use a snapshot schedule (a resource policy attached to the disk) with a retention period instead of manual snapshots.
  • Custom image: a bootable template made from a disk, snapshot or other image, to create identical VMs. Group versions with an image family so templates always pick the newest.
  • Machine image: captures the whole VM (all disks, configuration and metadata) for cloning or backup.
  1. Create — gcloud compute snapshots create or images create
  2. View — gcloud compute snapshots list / images list
  3. Delete — remove old ones; deleting a snapshot keeps the data that later snapshots still need

6. Operating data and networks, monitoring and logging

Securing and managing objects

  • Uniform bucket-level access: permissions come only from IAM, not per-object ACLs. Recommended.
  • Public access prevention stops a bucket being made public by mistake.
  • Signed URLs give time-limited access to one object to someone without a Google account.
  • Object Versioning keeps noncurrent versions when objects are overwritten or deleted. A retention policy (lockable) blocks deletion for a set period, for compliance.

Lifecycle management

Lifecycle rules act automatically on objects that match conditions such as age, storage class, number of newer versions or creation date. Actions: SetStorageClass (for example Standard to Coldline after 90 days) and Delete. Remember the minimum storage durations: a lifecycle move to Archive is free, but the object's total age (time in the earlier class counts) must reach 365 days before deletion, or an early-deletion charge applies for the remainder.

Data operations

TaskTool
Query BigQuery and estimate cost firstbq query --dry_run shows bytes scanned (on-demand pricing bills by bytes processed)
Query Cloud SQL, AlloyDB, SpannerConsole query tools or a SQL client; Bigtable with the cbt CLI; Firestore from the console or client libraries
Check a pipeline or query jobDataflow job page; bq ls -j and bq show -j for BigQuery jobs
Estimate storage costsGoogle Cloud Pricing Calculator
See the whole database fleet, health and recommendationsDatabase Center

7. IAM and service accounts

An IAM allow policy is attached to a resource (organization, folder, project or an individual resource such as a bucket). It holds bindings: a role, the principals who get it, and optionally a condition (for example only until a date, or only for resources with a given tag).

TaskCommand
View a project's policygcloud projects get-iam-policy PROJECT_ID
Grant a rolegcloud projects add-iam-policy-binding PROJECT_ID --member=group:ops@example.com --role=roles/compute.instanceAdmin.v1
Remove a grantgcloud projects remove-iam-policy-binding …
Find why someone has accessPolicy Analyzer / Policy Troubleshooter

Role types

TypeUse
Basic (Owner, Editor, Viewer)Very broad; avoid in production
PredefinedService-specific, maintained by Google; the normal choice
CustomYour own list of permissions when no predefined role is narrow enough. Created at organization or project level (not folder); you maintain them when services change.

Inheritance

Roles granted on a parent apply to all children; the effective access is the union. A role granted at the organization cannot be removed for one project by editing that project's policy. Use IAM deny policies for exceptions, and grant as low in the hierarchy as possible.

The full lesson continues in the interactive course on this page.

Sources and references

Unofficial practice course. Not affiliated with, endorsed by or sponsored by Google LLC. Google Cloud and Associate Cloud Engineer are trademarks of Google LLC. Check the current exam guide on Google Cloud's certification website before you book: Google updates the guide and product names regularly.

All courses · Cloud & DevOps - Free courses, certificate $5 | AMAADOR ACADEMY