How to pass a GCP certification? Cloud Architect & Data Engineer edition

8 min readOct 16, 2020

Business has moved to the cloud. And if it hasn’t it will. A cloud certification is currently one of the best assets to put on your resume and boost your career. Cloud Architect and Data Engineer are arguably two of the most valued roles and also a good point of departure before considering specialty certifications. In this article I’ll talk about how I prepared for both of them on one of the major cloud providers, Google Cloud Platform.

Disclaimer: this post is based solely on my experience and some of the advice I have been given when preparing for the tests, so it is of course no guarantee of success (but hopefully helps!).

Cloud Architect vs Data Engineer

Although the GCP services covered in the exam are potentially all, the particularities of each assessment will hint which are more relevant. The most straightforward way for you to gather the services you need to know by heart is to carefully read the exam guides. More specifically, these are the most significant products per test (as you would find them under the GCP console navigation menu):

Cloud Architect: Billing, IAM, Security, Compute, Storage, Networking, Operations, Tools.
Data Engineer: IAM, Storage, Databases, Big Data, Artificial Intelligence.

It is noticeable that the only two overlaps are IAM and storage. The former because keeping an adequate identity strategy is critical for every component of a cloud solution, and the latter because it is intertwined with every other GCP service, apart from being essential for cloud migrations (which is a recurrent topic for the architect role).

Once the outline has been cleared, it goes without saying that not every service has the same presence in the exam. In my experience, the gist of the exams narrowed down to the following GCP products:

Cloud Architect: Compute Engine, App Engine, Kubernetes Engine, Network Services.
Data Engineer: Dataproc, BigQuery, Bigtable, SQL.

It is also worth mentioning that I didn’t run into outside-GCP questions on my architect exam. On the contrary, the data engineer test included questions and/or options which required some knowledge of other technologies, most specifically Hadoop and Kafka.

Before sitting a GCP certification exam for the first time, I remember wondering whether there would be questions related to the gcloud command-line tool. Short answer yes. Elaborating a little bit more, there were some in both professional certifications, but significantly more in the associate one. On architect and data engineer, the focus was mainly on kubectl and bq, respectively.

Which one should you take? Well that depends largely on your interests and career perspective. My line of work is focused on data, so I can only speak from the perspective of a data engineer wondering whether to take the architect exam or not (but not the other way round) and maybe this is a matter of how specialised your data job is. However, I also think that you should always have the bigger picture and I strongly believe the cloud architect competences will help you become a better data engineer.

You perhaps want to pursue both certifications but are unsure about which one you should take first. The questions in the data exam are more specific. You will come across SQL and BigQuery in the two tests, but the questions in the data path will be much more in-depth. However, I would find it quite odd to find Kubernetes in the data engineer certification. That said, I take that the gained, more profound knowledge of the data exam will serve you well in a considerable share of the architect assessment.

Exam Format and Tips

50 multiple choice and multiple select questions to be completed in 2 hours. Pass or fail, you won’t get a percentage and you won’t know which answers were correct. And that’s pretty much it, I don’t really know how Google computes the pass or fail decision so just be prepared to ace all the questions.

It’s useful to know that you can mark questions for review and later revisit any of them before submitting the exam. Also and to prevent a potential heart attack, once you click submit you won’t get directly the result. First there is a screen in which you can fill in a survey about your experience so don’t freak out if you don’t see the outcome right away, you’ll get it in the next screen.

I am not going to talk about exam strategies because to each his own and to be fair I have never been the greatest at tests. If there is one thing I can say for sure though is be sure to read carefully the question statement and each one of the possible options. Many of the questions I’ve faced have really similar answers and the slightest detail tells apart a correct guess from an incorrect.

With respect to time management, in my experience I found 120 minutes to be enough to complete the exam. There is a countdown on the screen so you don’t need to worry about losing track of time (which is very nice because in none of my certifications I was allowed to bring my wristwatch into the testing room). Be aware that the clock doesn’t tick while you go from question to question, so take advantage of those few instants!

Online Resources

This is probably the section you were waiting for. There are many resources out there so it can be very easy to get sidetracked. Not surprisingly the first thing you may be interested in is actual exam questions. After all, other cloud providers such as AWS offer practice exams that you can purchase for additional training. This is unfortunately not the case for GCP. However, you do have some sample questions (not even 50 for each certification though) to familiarise yourself with the exam. According to what I saw in my examinations, these sample questions are somewhat similar to what you should expect to face.

Google offers a series of online courses divided by certification tracks on Coursera, which also include some quizzes. This is a great resource if you want to learn from scratch about any of the GCP services, since it gives an overall explanation of the purpose and main characteristics of each GCP product. However, Coursera won’t go as deep as the exam requires and I would have it as a high-level first step into your cloud journey. Regarding the quizzes, honestly the questions in the exam are nothing like those so don’t be mislead. All in all and as a general rule, I wouldn’t consider Coursera to be an infallible tool to pass the exam.

There are many other providers out there that offer online courses, from which I need to highlight Linux Academy. The content of the lessons is slightly more thorough than that of Coursera but the most remarkable trait of Linux Academy is undoubtedly the fact that their quizzes are much more realistic and similar to the real deal. One feature I found particularly useful, which I was recommended to make use of, is the deck of flash cards. They are user-created and shared with the community, meaning you can also create your own, and come in handy to keep track of everything covered in the exam curriculum. Also and while courses are tied to a subscription (both Coursera and Linux Academy, not taking into account free trials), you can access the flash cards on your free plan.

A considerable part of the exam involves design decisions which are mostly answered in the best practices and how-to guides sections of each GCP service documentation. Good examples to this are the best practices of Storage or the schema design for your Bigtable tables. I also highly recommend looking for all the decision trees Google has concerning choosing one or another of their products because such a decision is definitely going to be in the exam.

When lacking experience of diverse use case scenarios, the best approach to follow is to read about what other people have done (and then try those out yourself of course). And who to exemplify business solutions better than Google themselves! I highly recommend going through their solutions site. Even if not preparing any certification, it is a really good place to start designing your architectures.

To get a little bit more hands-on experience, use Qwiklabs. This is an online training site that will create a real GCP environment for you to use. Besides, you can take advantage of Google’s free GCP credit ($300) to start building your own solutions (there is really no better way to learn).

Another fantastic source of inspiration is the Google Cloud Blog. Apart from news about just out products and features, they regularly post articles about diverse applications of their GCP services. As a note on their brand-new products, you most likely won’t find questions related to newcomers or recent updates so don’t sweat over it.

With all honesty, these are all the resources I have used to prepare for the exams. I am sure there are many other superb tools but I cannot recommend anything that falls out of my experience.

Conclusion

Taking the decision to study for a certification is already a good choice for your career. Cloud Architect and Data Engineer belong to the GCP Professional set of certifications, for which Google recommends 3+ years industry experience including 1+ years on Google Cloud. Is this really necessary? In my humble opinion the answer is no. And by saying this I by no means want to reduce the value of these certifications (that’d be shooting myself in the foot), it’s just that I believe years of experience don’t have a perfect correlation with knowledge. Don’t get me wrong, your practical knowledge will be invaluable for both the exams but it is not foolproof. I am not the most experienced engineer out there but I do read and play around with technologies a lot. I think it all comes down to your interest in the subject and how hard you work.

Good luck and don’t hesitate to share your experiences and tips!