Google Cloud Professional Architect

Ever since working with a great team at the BBC, I’ve been noticing the momentum behind Google Cloud. I’ve been looking to do a certification for a while, so decided to go with Google’s Professional Cloud Architect.

Photo by Dewang Gupta

If you do some research on the certification, you’ll find words like “vast” describing the scope of the exam. They’re not wrong. From compute, storage and networking, through business requirements, capacity planning, SRE, regulatory compliance, containers, continuous deployment, even kubectl commands, this isn’t something you can study for straight out of the gate from coding bootcamp.

The range of topics and the layers of knowledge, from CIDR blocks to Continuous Deployment to cloud migration and hybrid connectivity, sets a high bar. That’s what makes it both tough and also respected. It takes a healthy and broad level of experience to tackle it. What’s nice is that experience is more than theoretical and more than rote product knowledge, and there are common-sense aspects in there too that probe real-world experience. Something I particularly like is that the answer isn’t always Google.

Preparing for the exam

The advice I was given was to go for the Coursera material. That stood me in good stead. Don’t expect perfection though: this stuff is changing all the time and there are a few bloopers and “human touches” in the content. My favourite is when the instructor’s Google Home starts talking to him in the background, closely followed by the time there’s a rustling sound, as if someone is monkeying around behind the camera, and the straight-faced, earnest instructor can’t quite hold back a lovely smile.

When it comes to the cloud, over-engineered perfection isn’t the name of the game and for me these little foibles add real warmth to what is otherwise a pretty intense process of learning. I listened to much of the content on 1.5x speed, partly to get through it and partly to stop my mind from wandering. There’s nothing quite like feeling like you‘re working to keep up to keep you focused. It’s important to say you won’t get everything from the course material. You’ll get good coverage of most areas, but it’s unlikely the course content alone will get you through.

I started with the Architecting with Google Cloud Platform Specialization and Preparing for the Google Cloud Professional Cloud Architect Exam. I’ve had hands-on experience with AWS and GCP at this point, but haven’t covered their breadth of services, so those courses broadened my horizon but, perhaps more importantly, underscored how many services there are and how much there is to know about each. Having completed the courses comfortably, and hit 80% on the practice exam, I felt I had a pretty good grasp of how under-prepared I was.

I’m strongest on compute: virtual machines, functions, Kubernetes and PaaS are all familiar to me, although I needed to get into the finer detail (e.g. how are storage throughput and network capacity affected by the number of cores on a Compute Engine instance?). I decided to round out the other areas covered by the exam.

I knew I had plenty to learn about the range of storage products, their different aims, use-cases, capacities, advantages and disadvantages, so I decided to do a few courses from the Data Engineering, Big Data, and Machine Learning on GCP Specialization. I also decided, because Kubernetes is a Leviathan with hidden depths, I’d do as much of the Architecting with Google Kubernetes Engine Specialization as I could before the exam. This gave me the detail I needed to answer one or two questions I might otherwise have had to make intelligent guesses for.

Going for the numbers

As with any certification exam, there’s diminishing returns to over-studying. You’ll likely only work in depth in a few areas and all areas will change over time, so knowing it all as it stands today doesn’t add much value. Knowing enough to pass across the board, staying up to date with most things and getting really good at a couple of things is my idea of pragmatic and practical. I booked the exam and dug in for some more study.

I now felt I had compute and storage covered, plus some more detail on GKE, which got me feeling more comfortable. I have a decent grasp of networking, and by now had a good idea of GCP’s take on VPCs and load balancing. I had touched on Stackdriver both theoretically and in practice, so held off on learning more there. I felt I still needed more detail though, so I went looking for blog posts like this one by Jean-Louis (JL) Marechaux and this one by sathish vj to get some leads on where to deepen my understanding. They were particularly effective in getting me more than adequately terrified about what I was facing.

Vanishing hope

In case you think I’m made of the stuff that causes impostor syndrome in innocent bystanders, someone sliding through with ease, I’d like to share with you the sense of foreboding I felt coming up to the exam. I hope that if you’re studying for PCA and all you see on this glistening Internet are what look to be smug people who sailed through under a light breeze, a glass of prosecco and strawberries in hand, pinkies out, shades on, that I can share a real moment which I hope will bring you solace.

It’s a tough exam. The range of things you realise you don’t know will grow faster than the number of things you do know. Much like life, you can’t “win” this one, but you can show up with the best of your efforts and experience. I ended up trawling through pages and pages of documentation, gleaning details, into the nigh, hoping that just a few what felt like a smattered collection of details I was gathering would get me over the bar — a bar looked very high.

I found a practice test to take, the night before the exam, on a website somewhere. I gave it a go, just to get an “exit poll” on what I’d learned ahead of exam day. It had 10 questions, two of which I recognised from the official practice exam. I scored 50%, including the two questions I already knew the answer to. It wasn’t good. With hindsight, I think the questions were open to interpretation, or maybe I was tired, but at the time I became increasingly concerned that I was about to face-plant.

Exam day

There’s no official pass mark for the exam and you get no feedback. Just a yes or no. It’s two hours and 50 questions. I was anxious. There was one timely bright spark: I’d watched this Simon Sinek talk in my worried state the night before. He explains that “anxious” and “excited” are physiologically similar. It turns out that just saying “I’m excited” rather than “I’m nervous” can materially improve your performance.

My exam was at the Pitman Training Centre in central Edinburgh. I arrived early, had my ID ready and they got me started. There’s something about these kinds of testing facilities where it feels like everything is running Windows Vista on a Pentium II and you’re never sure if the next screen is really going to load. I settled in. “I’m excited” I said out loud in my head.

I opted for a strategy of answering everything as a first guess, then review and re-review, gradually whittling down to those last few tough nut questions. The most important tip I can give you for this exam is also the most simple: read the question. And the answers. Pay attention to the language and don’t rush. The questions are well written. Don’t assume you know what’s being asked. Bookmark questions for review.

My first pass took just over an hour. In the next half hour I did a full review to eliminate the questions I was most confident in. I was left with about 15 to whittle down. I spent my last 20 minutes going over them, committing to answer one by one. I knew there’d be a few that would have be educated guesses. I wasn’t shooting for 100% but I knew I had to minimise the chance of slipping under the bar by a few points.

With two seconds left on the clock, I submitted my final answer and the exam was over. A feedback form later, I found myself staring at a white screen with writing on it, explaining what would happen next. After those two hours of intense concentration I was word-blind. I scanned the sentence back and forth, but saw no sign of a test result. Finally my eyes scanned up a little way and landed on a single word: Pass.

Over the rainbow

I’ve never been so pleased, so relieved and so thankful to see those four letters. For me achieving Professional Cloud Architect felt like stepping up to a genuine challenge. If you’re considering it, I’d certainly recommend it, but not for light entertainment.

Knowing what I do now I can say I’ve got a new level of respect for people who’ve done it. More than a learning experience, it’s the closest I’ve come to something that can assess those of us who identify as something like Architect or Tech Lead. There’s something both humbling and satisfying about putting yourself to the test and coming through.

When it comes to measuring yourself against the theory and practice of designing, building and operating modern cloud architectures, PCA is the highest and most relevant standard I’ve found. Thank you to Google Cloud and all the training course developers, it’s been a blast.

No certification post would be complete without some proof that I’m not just making this up to entertain you, so here’s the genuine article.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
David Carboni

David Carboni

Hands-on culture and techology. Work hard be kind. CTO at Policy in Practice (