Leanpub Header

Skip to main content

Topics

Google Cloud

From projects, the control plane and Cloud Run to networking, BigQuery, Firebase, migration and the certification exams, Google Cloud as told through the free samples of the Leanpub books about it.

Created from the free samples of 7 Leanpub books and the descriptions and tables of contents of 20 more books and courses, with every source cited.

Google Cloud, often still called Google Cloud Platform or GCP, is Google's public cloud: virtual machines, managed platforms for containers and functions, storage, databases, analytics and machine learning services, sold by use. One of the library's guides describes it as a collection of infrastructure and application services exposed through programming interfaces, all of which the web console, the command-line tools and automation use alike, and says the first question to ask of any of its services is which operational responsibilities it removes and which remain with the team.[1]

The platform is known for the things Google built for itself and then offered to others: Kubernetes, which it open-sourced and runs as a managed service; BigQuery, its data warehouse; and a family of serverless runtimes of which Cloud Run is the best regarded. A book on deploying applications lists five places a single web application can run on it,[2] and a beginner's guide puts the commercial case in the usual terms, capacity paid for as it is used by a company that does not yet know how much it will need.[3]

This article follows the subject from the project outward: where to begin, the history, how the platform is organised, the choices for running applications, networking and identity, storage, data and AI, migration, and the certification exams. Its sources are the free samples of the Leanpub books and courses listed at the end, with the descriptions and tables of contents of those whose samples were not available, and every claim that rests on a source is cited to it. Google Cloud is a thin shelf on Leanpub. Seven of its books have samples with real text, and many of the rest are short guides of 2026 known only from their descriptions, so this is a shorter article than Amazon's or Microsoft's platforms get.

Where to start

A newcomer has three introductions to choose from. Google Cloud A Complete Visual Guide (2026) is an illustrated course for new systems engineers and self-learners that runs in order from the resource hierarchy and access control through networking, compute and data to operations, around one small running example.[1] GCP Cloud Architecture (2020) is a short guide by several authors for people doing data engineering and data science, which begins with what the cloud costs and saves and with setting up an account and the command line.[3] Google Cloud Platform (2026) is a tour of the services with guides for choosing between them and comparisons with AWS.[8]

A developer with an application to ship should go to Deploying Node.js on GCP (2020), which deploys the same Node.js application five ways,[2] or, for mobile and web applications, to Firebase A Complete Visual Guide.[7] A candidate for the first certificate can start with Google Cloud Certified Associate Cloud Engineer Certification Guide 1, written for people with little or no experience of the platform.[4]

Engineers who already run workloads on Google Cloud are the audience for most of the rest: Mastering GCP for Web Applications: A Well-Architected Approach to Cloud Excellence for improving an existing environment,[5] Landing Zone in Google Cloud and VPC and Subnetwork Designing for foundations and networks,[9][10] The BigQuery Optimization Playbook for teams with a BigQuery bill,[18] and GCP Security & Governance for security engineers and those preparing for senior interviews.[13]

History

Google's cloud began in 2008 with App Engine, a platform that ran an application without exposing the servers under it. Virtual machines, in Compute Engine, and the BigQuery data warehouse followed in the early 2010s. In 2014 Google released Kubernetes, the container orchestrator derived from its internal systems, as open source, and offered it as a managed service the next year; Cloud Run, which runs a container on demand without a cluster to manage, arrived in 2019, and the machine learning services were gathered under the name Vertex AI in 2021.

The books carry these dates with them. The deployment guide of 2020 explains that Cloud Run is built on Knative, a project Google started for running serverless applications on Kubernetes, and records limits that applied at the time of writing, such as having no private access to a network.[2] An architect's study guide is offered free because, by its author's own note, it is out of date against the current exam.[6] A book of 2022 on taking deep learning to production includes training in the cloud and operations with Vertex AI,[24] while a service overview published in 2026 still lists the older names in its chapter on AI.[8] The newest title on the shelf prepares for an examination on the architecture of agent systems.[20]

How the platform is organised

Google Cloud A Complete Visual Guide starts from responsibility rather than product names. It arranges the services as a ladder: on a virtual machine the team manages the operating system, runtime and scaling; on a managed platform such as Cloud Run it manages the application, its identity and its limits while Google runs the machinery; on the managed Kubernetes service Google runs the control plane and the team still owns its workloads and policies. Its advice is to choose the highest rung that fits the workload without forcing an awkward design, and its analogy is transport, a leased car, a taxi, a delivery service. The rungs are not rigid, it adds: managed Kubernetes removes the administration of the control plane and still asks for more platform knowledge than a managed container service, and a database service that takes backups automatically leaves the customer to set the policy and test the restoration. The same chapter separates the control plane, through which resources are created and configured, from the data plane that serves traffic, a distinction it calls practical during incidents: a failed deployment points to permissions, quotas or configuration, while errors after a successful one point to routing, health or capacity, and the existence of a resource is no evidence that the application is correct. On responsibility it gives Google's own formulation, shared fate, under which the provider supplies secure defaults and guidance and not merely a boundary, and then a rule of its own: never infer a safety property from the word managed. A service that adds instances as traffic rises, for instance, helps only if the application starts correctly and its dependencies can bear the load.[1]

The guide's running example, a small internal inventory service, is used to trace one request from end to end: a name is resolved, a load balancer accepts the connection and picks a healthy backend, a compute service runs the code, the application authenticates to a data service, the data is read or changed, and logs, metrics and traces record what happened. Each step, it points out, is a design decision that can be got wrong, a stale name record, an identity with more privilege than it needs, a database that is available and overloaded, logs that exist and omit what a diagnosis requires; its comparison for the two planes is a restaurant, where changing the menu is one kind of work and serving a meal another.[1]

The unit of organisation is the project, and projects sit in folders under an organisation. Mastering GCP for Web Applications: A Well-Architected Approach to Cloud Excellence recommends separate projects by purpose: one for organisation-level administration that holds no workloads, production projects under strict change control, non-production projects for development and testing, and a logging project that other projects write to and ordinary users cannot alter, with write-once storage suggested for the most important logs.[5] Landing Zone in Google Cloud (2026) follows a fictional financial company through the building of such a foundation from design to day-to-day operation.[9] Day to day, the platform is driven from the command line: GCP Cloud Architecture introduces the tool, whose commands are grouped by product, the named configurations that let one person move between projects and accounts, and the flags that name an account and project explicitly whatever configuration is active.[3]

Running applications

Deploying Node.js on GCP takes one sample application to App Engine, Cloud Functions, Cloud Run, Compute Engine and Kubernetes Engine in turn, with the same headings for each, deployment, debugging, builds, logging, security and when to choose it. Its sampled chapter is on Cloud Run. The advantages it lists are deployment of any container with one command, payment only for resources used, scaling that is fast because no virtual machine has to start, and up to eighty concurrent requests to an instance where a function instance takes one. The disadvantages are cold starts, the cost of a service that is called continuously and would be cheaper on something always on, the requirement that a container hold no state and do no work after its request ends, and a limit on how long a request may run. The service comes in two forms, one fully managed and one for running on an organisation's own clusters, and the chapter keeps to the first. It adds advice on sizing memory from concurrency, working an example in which eighty simultaneous requests of a megabyte each call for half a gigabyte, notes that lowering the concurrency limit to one, as when each request needs most of the processor, makes spikes of traffic start many instances from cold, and warns that idle instances may be kept for an undisclosed time, so an application should be built to start quickly rather than to rely on them. Each deployed service receives its own secure address, custom domains can be mapped to it, and a new revision can be rolled out gradually, with traffic split between revisions, or rolled back. Typical uses, by the book's list, are websites and web applications backed by a managed database, interfaces for mobile applications, and small pieces of code that join other services together. Access follows the project's roles: owners and editors can deploy and invoke a service, and only owners and the service's administrators can make it public.[2]

Kubernetes has two short books here. Hands-on Kubernetes with Google Cloud Platform (2024) is a quick overview of Kubernetes basics whose title names the platform,[15] and Service Mesh Troubleshooting Guide on GCP (2026) is a method for diagnosing a service mesh on it.[14]

Firebase is the route many application developers take. Firebase A Complete Visual Guide (2026) defines it as a managed application platform on Google Cloud in which client applications use identity, data and file services directly when policy enforced by the server allows, and insists that the engineering task is therefore not to remove the backend but to divide responsibility correctly among untrusted clients, managed services, trusted server code and human operators. Every Firebase project, it notes, is a Google Cloud project underneath, with the same access controls, locations and billing, and its second chapter draws the consequences: separate projects for separate environments, the location of resources treated as an architectural decision, and budget alerts understood as signals rather than brakes.[7] Firebase in Production, Without Surprises (2026) is a short list of production pitfalls gathered over thirteen projects, each with its symptom, cause and fix.[16]

Networking and identity

Google Cloud's network model differs from the other large providers', and VPC and Subnetwork Designing (2026) gives a chapter to the differences, among them that a virtual private cloud is global while its subnets are regional, before it turns to subnet planning, topology, firewall rules and connectivity.[10] PNA and PSC in Google Cloud explains two ways of reaching services privately through a story about a fictional company and a short test for choosing between them.[11]

Access is governed by identity, for software as well as for people. The visual guide's trace of a single request has the application authenticate to its database with a workload identity, one that represents software rather than a person;[1] Workload Identity is a short guide to exactly that mechanism for workloads on Kubernetes and for pipelines.[12] Mastering GCP for Web Applications: A Well-Architected Approach to Cloud Excellence covers the human side, groups, two-step verification, custom roles and the protection of the organisation's administrator account.[5] GCP Security & Governance (2026) is a senior-level handout on governing a whole estate, from organisation policy to security posture,[13] and DevSecOps Tools in Cloud Platform treats Google Cloud's security features briefly beside those of two other providers.[27]

Storage, data and AI

Cloud Storage, the object store, is the most fully documented service on the shelf. The sampled chapter of Unofficial Google Cloud Certified Professional Cloud Architect Study Guide sets out its storage classes, standard for data in frequent use and three cheaper classes with minimum storage periods of thirty, ninety and three hundred and sixty-five days, all with the same low latency; the rules that move or delete objects as they age, each one action triggered by one or more conditions; and the meaning of the service's durability figure, which the author renders as the loss of one object in ten thousand years out of ten million stored, with the caution that no such figure protects a bucket confined to one region from the loss of that region.[6] The course Deep Dives: Google Cloud Storage stays on the one service, from buckets and access control to lifecycle, retention and pricing.[17]

For analytics, The BigQuery Optimization Playbook (2026) ranks thirty-three patterns for cutting the cost of BigQuery, starting with those that reduce the data a query scans, and discusses the choice between paying by the query and reserving capacity.[18] Full-Stack Data Engineering with Node.js reaches the same services from Node.js,[26] and Cloud Computing for Data Analysis introduces Google Cloud beside two other clouds for people doing data science.[25] On machine learning, Deep Learning in Production uses Google Cloud as the place where its models are trained and served.[24] Dialogflow CX (2026) covers the platform's tool for building conversational agents, with a banking example,[19] and Google Cloud Professional Agentic Architect links the published requirements of the agent architecture exam to design decisions about tools, retrieval, memory, evaluation and security.[20]

Migration

Two companion guides of 2026 treat a move to Google Cloud as a fixed sequence. Migrating Stateless Workloads to Google Cloud has twelve steps, from assessment through deployment and cutover to cleanup, with a rollback plan,[22] and Migrating Stateful Workloads to Google Cloud has thirteen, for databases, file storage and session stores, ending with the decommissioning of the source and including a walk through the move of one database to the managed database service.[23] Whether to move at all is the question GCP Cloud Architecture puts first. It groups the costs and benefits under four heads, cost, security and compliance, data management and scalability. Its answer is that the cloud suits a company that cannot yet size its needs, that at some point the balance of cost may shift back, and that the centralisation of security and of access logs is a benefit in its own right, which it illustrates with a laptop full of health data stolen from a car.[3]

Certification

Half of the shelf's samples come from examination guides. Google Cloud Certified Associate Cloud Engineer Certification Guide 1 (2026) is the first of three volumes for the Associate Cloud Engineer exam; its author describes that exam as weighted towards running systems and the command line, the architect's exam as weighted towards business analysis and trade-offs, and both as requiring that the candidate has built things by hand, which is why the course is organised around challenge labs; many candidates, he says, treat the first as a stepping stone to the second. He wants candidates to spend almost no time memorising and to pass because they understand the platform well enough to reason through a scenario; the volume sampled covers how the platform is laid out, setting up an account and projects, and working from both the console and the command line.[4] Google Cloud - Associate Cloud Engineer - 250+ Exam Questions is a bank of practice questions for the same exam.[21] Unofficial Google Cloud Certified Professional Cloud Architect Study Guide (2021) condenses the Professional Cloud Architect syllabus to what its author thinks the exam asks, with its case studies.[6] The security handout and the agent architecture guide, described above, are each tied to an exam as well.[13][20]

Questions about Google Cloud

Which Google Cloud book should I start with?

A newcomer can begin with Google Cloud A Complete Visual Guide (2026), an illustrated course for new systems engineers and self-learners that runs in order from the resource hierarchy and access control through networking, compute and data to operations.[1] A developer with an application to ship should go to Deploying Node.js on GCP (2020), which deploys the same Node.js application five ways.[2] A candidate for the first certificate can start with Google Cloud Certified Associate Cloud Engineer Certification Guide 1, written for people with little or no experience of the platform.[4]

Should I use a virtual machine, Cloud Run or Kubernetes on Google Cloud?

Google Cloud A Complete Visual Guide arranges the services as a ladder: on a virtual machine the team manages the operating system, runtime and scaling, on a managed platform such as Cloud Run it manages the application, its identity and its limits, and on the managed Kubernetes service Google runs the control plane while the team still owns its workloads and policies.[1] Its advice is to choose the highest rung that fits the workload without forcing an awkward design.[1]

What is Cloud Run good for and what are its drawbacks?

Deploying Node.js on GCP (2020) lists the advantages of Cloud Run as deployment of any container with one command, payment only for resources used, and scaling that is fast because no virtual machine has to start.[2] The disadvantages are cold starts, the cost of a service that is called continuously and would be cheaper on something always on, the requirement that a container hold no state, and a limit on how long a request may run.[2] Typical uses, by the book's list, are websites and web applications backed by a managed database, interfaces for mobile applications, and small pieces of code that join other services together.[2]

Is Firebase part of Google Cloud?

Firebase A Complete Visual Guide (2026) defines Firebase as a managed application platform on Google Cloud in which client applications use identity, data and file services directly when policy enforced by the server allows.[7] Every Firebase project, it notes, is a Google Cloud project underneath, with the same access controls, locations and billing.[7]

What are the storage classes in Cloud Storage?

Unofficial Google Cloud Certified Professional Cloud Architect Study Guide (2021) sets them out: standard for data in frequent use, and three cheaper classes with minimum storage periods of thirty, ninety and three hundred and sixty-five days, all with the same low latency.[6] Rules can move or delete objects as they age, each one action triggered by one or more conditions.[6]

What is the difference between the Associate Cloud Engineer and Professional Cloud Architect exams?

The author of Google Cloud Certified Associate Cloud Engineer Certification Guide 1 (2026) describes the Associate Cloud Engineer exam as weighted towards running systems and the command line, and the architect's exam as weighted towards business analysis and trade-offs.[4] Both require that the candidate has built things by hand, and many candidates, the author says, treat the first as a stepping stone to the second.[4]

Sources

Every source is a Leanpub book or course. The note after each says which part of it was read; where no sample was available, only the public description and table of contents were used.

  1. 1.
    Google Cloud A Complete Visual Guide, Reiner Braun: A Complete Practical Guide to Foundations, Architecture, Delivery, and Operations Free sample: the introduction and chapter one, on the service-management ladder, control and data planes, tracing a request and shared responsibility. 2026. Free sample: PDF · EPUB
  2. 2.
    Deploying Node.js on GCP, Jonathan Lin: A Comprehensive Guide to Deploying Node.js on Google Cloud Platform Free sample: the chapter on Cloud Run, its advantages, disadvantages, deployment, concurrency, cold starts and access control. 2020. Free sample: PDF
  3. 3.
    GCP Cloud Architecture, Noah Gift, Andrew Nguyen, Alfredo Deza, Michael Vierling Free sample: chapter one, on the costs and benefits of the cloud, creating a project and using the command-line tool. 2020. Free sample: PDF · EPUB
  4. 4.
    Google Cloud Certified Associate Cloud Engineer Certification Guide 1, Bolakale Aremu: Learn with challenge labs, assessment tests and practice exams to build innovations and real-life experiences Free sample: the contents and the introduction to the course and the exam. 2026. Free sample: PDF · EPUB
  5. 5.
    Mastering GCP for Web Applications: A Well-Architected Approach to Cloud Excellence, Chinmoy Mukherjee Free sample: the introduction and the chapters on project structure and identity. 2026. Free sample: PDF
  6. 6.
    Unofficial Google Cloud Certified Professional Cloud Architect Study Guide, Jonathan Lin: The Missing Lecture Notes for Passing the Professional Cloud Architect Exam Free sample: the chapter on storage. 2021. Free sample: PDF
  7. 7.
    Firebase A Complete Visual Guide, Reiner Braun: Architecture, Security, Data, Compute, Delivery, and Operations — Chapters 1–28 Free sample: the contents and the opening of the main text on what Firebase is. 2026. Free sample: PDF · EPUB
  8. 8.
    Google Cloud Platform, Sudhanshu Jaiswal: A Walkthrough No sample: from the description and table of contents. 2026.
  9. 9.
    Landing Zone in Google Cloud, Sudhanshu Jaiswal: A Practical Guide to Cloud Excellence No sample: from the description and table of contents. 2026.
  10. 10.
    VPC and Subnetwork Designing, Sudhanshu Jaiswal: in GCP No sample: from the description and table of contents. 2026.
  11. 11.
    PNA and PSC in Google Cloud, Sudhanshu Jaiswal: A Guide to Private Connectivity No sample: from the description and table of contents. 2026.
  12. 12.
    Workload Identity, Sudhanshu Jaiswal: in GCP No sample: from the description and table of contents. 2026.
  13. 13.
    GCP Security & Governance, Sudhanshu Jaiswal: Senior-Level Study Handout No sample: from the description and table of contents. 2026.
  14. 14.
    Service Mesh Troubleshooting Guide on GCP, Sudhanshu Jaiswal: A Practical Companion for Cloud Engineers No sample: from the description and table of contents. 2026.
  15. 15.
    Hands-on Kubernetes with Google Cloud Platform, Pablo Maldonado No sample: from the description and table of contents, which describe Kubernetes generally. 2024.
  16. 16.
    Firebase in Production, Without Surprises, Leandro Salerno: Pitfalls measured on 13 real projects No sample: from the description alone; the table of contents is empty. 2026.
  17. 17.
    Deep Dives: Google Cloud StorageCourse, deepdives.eu: Deep level learning of Google Cloud Storage No sample: from the course description and lesson list; the index records no publication date.
  18. 18.
    The BigQuery Optimization Playbook, Nitin Gandhi: Patterns and Practices for Cost and Performance (2026 Edition) No sample: from the description and table of contents. 2026.
  19. 19.
    Dialogflow CX, Sudhanshu Jaiswal: Building Next-Gen Conversational AI No sample: from the description and table of contents. 2026.
  20. 20.
    Google Cloud Professional Agentic Architect, Daniel Vaughan: Principles, Products and Practice for Passing the PAA Exam No sample: from the description and table of contents. 2026.
  21. 21.
    Google Cloud - Associate Cloud Engineer - 250+ Exam Questions, Luis No sample: from the description alone. 2025.
  22. 22.
    Migrating Stateless Workloads to Google Cloud, Sudhanshu Jaiswal No sample: from the description and table of contents. 2026.
  23. 23.
    Migrating Stateful Workloads to Google Cloud, Sudhanshu Jaiswal No sample: from the description and table of contents. 2026.
  24. 24.
    Deep Learning in Production, Sergios Karagiannakos From the description and table of contents. 2022. Free sample: PDF
  25. 25.
    Cloud Computing for Data Analysis, Noah Gift: The missing semester of Data Science No sample: from the description and table of contents. 2021.
  26. 26.
    Full-Stack Data Engineering with Node.js, Kevin Languedoc: How to build a Data Engineering and Business Intelligence Stack using Node.js and Npm modules From the description and table of contents. 2024. Free sample: PDF
  27. 27.
    DevSecOps Tools in Cloud Platform, Joseph Thachil George: DevSecOps From the description and table of contents. 2022. Free sample: PDF

Further reading

Other books in the library on the same ground, not cited above.

  • The CloudStorageMaven Guide, Emmanouil Gkatziouras A guide to a tool that keeps Maven repositories on cloud storage, Google's included. 2019.
  • Helm Troubleshooting Guide, Sudhanshu Jaiswal Diagnosing Helm problems on Kubernetes, with a note on credentials for Google's managed service. 2026.
  • Apache Airflow, Rohit Jain: Quick Start Guide A guide to Apache Airflow whose cloud chapters use AWS and compare its managed service with Google's. 2026. Free sample: PDF
  • Cloud Computing: From Fundamentals to Advanced Concepts, Anshuman Mishra A vendor-neutral university textbook on cloud computing that names Google Cloud among its examples. 2026. Free sample: EPUB

This is an article created from the free samples of 7 Leanpub books and the descriptions and tables of contents of 20 more books and courses. We're building these articles to provide you, the Leanpub reader, with a good jumping-off point to discover more Leanpub books that are interesting and relevant to you. We never use material which is paid-only when building these articles: we only use material which is already free on Leanpub. However, Leanpub authors can opt out from inclusion in articles from their book settings page.

Updated November 2026.

In the Leanpub store: 49 books · 2 courses in the Google Cloud Platform category

All topics