How I fail the GCP Professional Cloud Architect certification exam

Aurélie Vache - Jan 4 '21 - - Dev Community

First of all

it's not an easy article to write, but I think it can be useful for people who want to pass this certification exam.

Today I try to pass the exam for the GCP Professional Cloud Architect certification, and I failed.

It's not a good feeling. I'm disappointed with myself, but I know that I learned a lot during my preparation and that's what matters.

Preparation

Before to try to pass this exam, I prepared myself during approximately one month.

The first thing I can recommend you is to ask an on-demand training for the GCP certificate you want to have.

I follow the whole course and quests in qwiklabs in my side (and an Anthos training):

Alt Text

This on-demand training allows me to improve my knowledges but it's not the only thing you should do.

During the exam you will have (not easy ^^) 50 questions, in which you need to imagine you in the role of the context given in the question or in one of the 3 official uses cases.

I had about 10 questions regarding the use cases.

Even if you will have the use case, on the right of your screen, during the exam, it is advisable to know the three use cases well and to know the solutions that would be considered.

I also watch a lot of videos on Youtube about preparation, cases studies and with sample questions.

And, of course, I passed the sample questions form.

Spoiler alert: the real questions are harder ;-).

Useful links

Questions samples

There are 3 case studies for this exam and the advice is to learn, learn, and know nearly by heart the requirements, the business needs ... in order to answer the 10-15 questions of the exam in relation to these cases studies.
For example TerramEarth with its 200,000 vehicles if they want to stream their data, we already know that we can eliminate BigQuery from the possible responses because the existing quota, at the present time, is 100,000 streaming inserts per second :-).

When we read "cost-efffective", we can immediately say that Dataproc (hadoop and spark managed in GCP) can be expensive, so we can eliminate it from the start as well, we will have to prefer a solution that can be less "hype" but more cost effective :-).

Feedbacks

This exam was not easy because when I read questions, when I have to think about an architecture, make decisions, choose the right solution, I immediately have the reflex to take a paper a pen and I sketch, I draw , I doodle, or I highlight the important words in the question, and there, I was not allowed to have all of that (whereas in real life if ^^).
So not allowed to read the question aloud, to highlight important words and to skit, it can increase the level of difficulty of comprehension and give the correct answers.

The thing that I find unfortunate is not knowing exactly which questions I missed and their solution. I remember a little where I had hesitations but at the end of the exam we know if we pass or fail, we do not have detailed results, and I find that very, very unfortunate.

There are questions with acronyms that I personally didn't know before preparing for the exam: HIPAA, PCI DSS ..., these are regulations and standards, so my advice is to familiarize yourself with these acronyms:

The important thing to know is that Google guarantees that the products covered by this agreement meet the requirements of the HIPAA law and comply with our ISO / IEC 27001, 27017 and 27018 certifications, as well as the SOC 2 report.

Advices :

  • Follow the qwiklabs (or coursera) course and labs

Even if there are bugs in some labs (I have found several), the course is quite complete (too much?).
On the other hand there are three courses concerning GKE and Kubernetes whereas I only had 5 questions out of 50 during the exam, it is suddenly a little light level questions compared to the number of courses to follow.

Personally I recommend going more smoothly than me and taking you more than a month to pass the courses and labs, and give your brain time to store everything :-D.

  • Learn well almost by heart the cases studies of Mountkirk games, Terramearth and Dress4Win (current architecture, business needs ... and know the possible technical solutions)

  • Read the questions slowly and several times (yes with stress we tend to want to quickly go from question to question), suddenly we read too quickly, and since each word is important, we can skip the most important word and in fact answer the question incorrectly).
    You have the time to answer all the questions and to review all of them at the end :-).

So, I failed the exam, I am not certified but does that mean that:

  • I did not learn anything ? The answer is NO
  • I do not know anything about it ? the answer is also NO. Preparing myself allowed me to consolidate my knowledge and to dig, to dig options, configurations, features that I did not know yet.
  • Do I need to improve? Definitely, YES ^^

What is a shame is that we know at the end of the exam whether we fail or succeed but we do not know his score :-(.
I want to improve my knowledges, so why it should be not possible to know the score ?

I took the CKAD last year and it's a little unsettling at the start because the CKAD is based on practice, you type commands to answer questions, it requires handling on a daily basis ...
And there you have to answer a MCQ, understand what the words mean, know a slew of GCP services, every nook and cranny, features and configuration and know the three case studies.

The good news

During my preparation, I took a lot of notes so during the coming weeks / months, a new series of sketchnotes will emerge: "Understanding GCP in a visual way". I hope you'll like it :-).

I also started to sketch notes with tree decisions, that can help you in your GCP journey :-):

Alt Text

What's next?

Try again in several weeks ;-).

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .