Intro

I recently sat and passed the Google Cloud Professional Cloud Architect Exam. This post covers my thoughts on the exam, the key topics and the learning resources I used.

The Professional Cloud Architect Exam is the ‘keystone’ Google Cloud exam. It covers the whole spectrum of Google Cloud products, bridging together many of the concepts covered in the more specialised certifications.

Google defines a Cloud Architect as:

"… [an] individual [who] designs, develops and manages robust, secure, scalable, highly available, and dynamic solutions to drive business objectives."

"…experienced in software development methodologies and approaches including multi-tiered distributed applications which span multi-cloud or hybrid environments…"

"…be proficient in all aspects of enterprise cloud strategy, solution design, and architectural best practices."

Google Cloud

This very broad job scope is reflected by the comprehensive knowledge of Google Cloud expected for the exam.

As with the other Google Cloud certifications, the exam follows the common format of 60 multiple choice questions in 2 hours. However, the exam also contains questions (12 in my case) based on any of four client case studies that you have access to and can study before the exam. The exam can be taken at a testing site or remotely proctored.

Disclaimer/my previous cloud experience

Your mileage with this exam will vary depending on your previous cloud and domain experience.

I have just over 12 months of experience working with Google Cloud on almost a daily basis.

Additionally, I come from a data science and data engineering background and my daily work revolves around Google’s database, data processing and analytics/ML tools such as Pub/Sub, Dataflow, BigQuery, Vertex AI etc. Many of the exam questions cover networking, security and DevOps concepts which were new to me but may be easier to pick up for practitioners already working in those domains.

My thoughts on the exam

This is my third Google Cloud exam. I have previously taken the Google Cloud Professional Data Engineer and Google Cloud Professional Machine Learning Engineer certifications. For me, the Cloud Architect exam was definitely the hardest of the certifications I have completed so far, for three main reasons:

  1. Broad Scope
    • Almost every Google Cloud product is in scope for the exam, and there are a lot of them!
    • The ‘specialist’ Google exams, such as the Professional Data Engineer, focus on a relatively small and defined subset of products. For example, for the Professional Data Engineer exam questions are largely limited to databases and data processing tools. This allows you to clearly focus your revision and get to know each product in detail. With the Cloud Architect exam, any product is fair game, from networking, project administration and DevOps to cost control, databases and machine learning. Therefore, you have to spread your revision time thinly across all these domains.
  2. Low experience with many tools
    • Following on from point one, due to the sheer volume of tools, it is unlikely that you have hands-on experience with most of the tools or domain expertise to understand their nuances. For example, I rarely work with or have to set up networking, VPNs or administer projects. Normally other members of the team are responsible for these functions. I found it quite challenging to learn the many concepts across all these domains, let alone confidently answer questions on them.
    • Having hands-on experience with the tools is a major advantage when answering exam questions on them.
  3. Subjective answers
    • Being a Cloud Architect is all about designing the right solution for the client, given their requirements and constraints. This means the right solution for the client may not necessarily be the most straightforward or even Google’s recommended solution.
    • For many of the questions, multiple answers will be ‘correct’ and you will have to identify which is the ‘most correct’ choice which can be subjective at times.
    • From my experience with the other Google exams, I find Google tends to use them as a marketing exercise for their managed products. Generally, you could get away with looking for ‘key’ words such as ‘streaming’ and ‘analytics’ and then automatically go to the answer section and select ‘PubSub, Dataflow, BigQuery’. With the Cloud Architect exam, you cannot be so naive! Most of the multiple-choice answers will be plausible, however, you really need to read the question carefully to identify whether the preferred Google product is appropriate or not for this particular situation.

Most important topics

You should have a good grasp of database, data processing, administration, deployment and networking tools before attempting this exam. However, I would say the exam is heavily skewed towards the following topics which you should know in detail before attempting the exam:

  • Google Kubernetes Engine (GKE) and Kubernetes cluster management
  • Compute Engine. Understand the process of backing up VMs, startup/shutdown scripts, security considerations, migrations from on-premise and networking settings. Many client case studies involve moving legacy infrastructure to the cloud. This normally means infrastructure running on an on-prem data centre. Therefore, there are many questions about how to migrate on-prem workloads to cloud VMs as a first step of moving to the cloud
  • App Engine. App Engine featured surprisingly heavily in the exam. You should know and understand the differences between the standard and flexible environment options.
  • DevOps tools (e.g. cloudbuild, container registry, binary authorisation). You should understand the DevOps workflows on Google cloud, such as building, securing and deploying containers in an automated fashion.
  • Security (IAM, Service Accounts, VPN products, Secrets management). You should understand basic security best practices for IAM such as the principle of least privilege as well as the various VPN products and methods for securing and managing secrets in your applications. Many of the case studies involve moving data from on-prem which requires a VPN connection.
  • Networking tools. Similar to the security section, you should understand what VPC and subnets are, their features and how they are used to secure your environments. This is a very important topic in the exam.
  • Object Storage (e.g. GCS). Understand object lifecycle management, cost-saving and latency considerations. This also includes knowledge of products such as Cloud CDN for serving content closer to your end-users around the world.
  • Compliance and regulations. Some questions were testing your understanding of major data protection regulations such as GDPR, SOX, HIPAA and PCI. Make sure you understand to which type of data these regulations cover and in which geographical regions.

In addition to these topics, I received 12 questions from 2 of the 4 case studies (EHR Health and TerramEarth). In all honesty, these questions were very similar to the other 38 general questions and generally could be answered without even reading the case study. However, I would recommend being somewhat familiar with the case studies beforehand as it will save time when reading the question as you are already aware of the context. Before taking the exam, you should have an idea in your head of the main products which could be appropriate (and more importantly not appropriate) for the case study client. This will help you eliminate wrong multiple choice answers immediately.

The exam was very light on machine learning tools - I don’t think I received any questions related to it, other than brief references to the ML APIs.

Learning resource

Listed below are the main resources I used for studying for the exam:

  • Google Documentation - Usefulness rating: 10/10
    • As with any Google certification, I always recommend reading through the ‘Concepts’ pages of the documentation for each tool. These pages provide a very good overview of the tool and its use cases in more than enough detail for the exam. I recommend for this exam really getting to grips with the Networking productsGKECompute Engine  and App Engine  documentation.
  • Qwicklabs - Usefulness rating: 7/10
    • I completed the Cloud Architect learning path . It is really important to get hands-on experience with as many tools as possible. I find that the hands-on labs solidify the concepts from the documentation and help you understand the nuances of each tool particularly if you do not use the tools daily. My usefulness rating is slightly less than the Google Documentation as the Architect learning path does not cover all aspects of the exam syllabus.
  • A Cloud Guru - Usefulness rating: 7/10
    • I signed up for a week free trial to ‘A Cloud Guru’ and took the Google Certified Professional Cloud Architect 2020  course. This was a good resource and contains a high-quality practice test with 50 exam questions. However, unfortunately, it has not been updated for the new syllabus and there is some outdated content and no questions on the newer case studies.
  • Priyanka Vergadia’s (aka The Cloud Girl ) sketch notes - Usefulness rating 6/10
    • Priyanka is a Google Cloud Developer Advocate and produces quality content explaining Google Cloud tools and concepts. In particular, she has created several ‘sketch notes’ which summarise the key components of many Google Cloud tools and comparisons between them. These one page summaries are great for refreshing your knowledge on each tool.
    • You can also see my other Top 10 Google Cloud Resources Article  for similar resources.
  • Whizlabs Practice Exam Questions - Usefulness rating: 8/10
    • One thing that was lacking in my preparation was good quality practice exam questions. The exam syllabus has been updated earlier this year and very few of the online courses and practice questions have been updated.
    • Google provides around 20 sample practice questions for free but I was unable to find many other quality free resources for practice questions.
    • The Cloud Guru course had a good set of 50 questions included in the free trial, however, they were a little outdated and did not cover questions on the new case studies.
    • With a severe lack of practice questions and in a crisis of confidence on the exam day, I resorted to paying for the Whizlabs exam questions  to help boost my confidence and power through some practice questions. After paying (£29.95) for the tests you get access to over 250 exam questions. On the whole, the questions are of good quality and similar to the questions faced in the final exam. However, I found the case study questions a little lacking. In the absence of good quality free resources, I would still recommend these practice questions as they were very useful for my preparation on the day.

Parting thoughts and recommendations

The Google Cloud Professional Cloud Architect Exam is challenging but highly rewarding. I would recommend Google Cloud practitioners from all backgrounds endeavour to work towards it during their careers. Since starting my studying for the exam, I have been able to have much better conversations with colleagues and clients. With any cloud technology, there are often multiple tools that can be used to complete the job. Having comprehensive knowledge of the wide spectrum of Google Cloud tools and their pros and cons is invaluable for providing value to your clients or business.

The exam helps hone your critical thinking skills to read and interpret client briefs and design better solutions tailored to the client’s needs. It is very important to be able to consider all possible options and not just be biased towards the tools you have the most experience in.

Exam technique is key for this exam. Multiple answers for a question will be ‘correct’. You need to be extremely careful to read the question properly to identify any constraints which eliminate certain tools, even if they are Google’s recommended product. Use the process of elimination to narrow down your choices.

I would not recommend completing the Cloud Architect exam as your first exam. The exam is challenging and requires a broad understanding of Google Cloud and cloud technologies. You should look into completing the most relevant ‘specialised’ exam for you first. For example, if you are a data engineer complete the Data Engineering certification first, if you are a data scientist complete the machine learning engineering exam and if you are a DevOps engineer complete the DevOps certification etc.. The specialist exams will provide you with the most value to your day-to-day role and will likely be easier for you to complete as you already possess the domain knowledge.

Other Relevant Articles

You can view my Professional Cloud Architect Certificate here

Good luck in your preparation!

Further Reading