The Blog

Corey Quinn’s AWS Beta Certification Exam Report

Calendar Icon 03.10.2021
aws-section-divider aws-section-divider

With my bizarre re:Invent lounge pass (read as: Cloud Practitioner Certification) expiring later this year, it behooves me to take another AWS certification exam to keep the lounge goodies flowing.

This year I opted for the “beta” Sysops Administrator Associate exam.

I give it the scare quotes around “beta” because—when you’re charging people money (in this case $75)—you don’t get to weasel out of a sub-par experience with that label.

Let me begin by saying I don’t know if I passed, nor will I for the next 13 weeks. This isn’t about me being particularly gruntled or disgruntled with the results of the exam. Instead, I want to talk about the exam process itself.

As my business partner discussed in some detail back when he took Cloud Practitioner, there are some sharp edges around the Pearson VUE testing experience.

A number of those have been softened, and overall the experience was positive.

The test itself

Let’s begin with the test itself, and the content therein.

I did no preparation for the exam, unless you count the last decade of my life fighting with AWS. The questions were aligned heavily with “Old AWS” in the sense of focusing on EC2, S3, and RDS. There were more questions about autoscaling groups than actual autoscaling groups I’ve configured in my career! None of this is new.

What was new for me was that this is the first certification exam that has a “lab” component. Specifically, there are three scenarios you’re given, and then granted “full” access to the AWS console or (theoretically) the AWS CLI.

When I say “the console,” I want to be clear here: It’s not some mocked-up testing environment version; it’s the real thing.

This got me into some trouble when I clicked a link that apparently led to the AWS docs site—the entire thing froze in a solid white screen that required a bit of jiggery-pokery to get back to the environment. Had I known this going in, I’d have done something while I was in the environment to run up Pearson’s AWS bill into the stratosphere just out of general principle.

The lab experience itself was fantastic, and more exams should focus on this kind of testing. There is of COURSE more than one way to get to a satisfactory outcome for the scenario, and empowering students to demonstrate mastery of those various ways is a heady feeling.

In fact, the multiple choice trivia questions just felt sad and diminished even further in comparison. Working with AWS isn’t about memorizing trivia like “the order of fields in VPC Flow Logs,” it’s about understanding how to find the answers quickly.

I’d be in favor of a pure-labs certification experience that was open book; I promise you, nobody is doing this stuff from memory in reality.

And if the test isn’t designed to reflect knowledge of how to work with AWS in production environments, then what’s the point of having the certification at all?

The experience

The “online proctoring” experience grates.

Maybe it’s because I’ve been out of school for 20 years, maybe it’s because of my anti-authority streak. Whatever the reason, being told to raise my hands to prove I’m not wearing a watch in my own home is disgusting to me.

I want to point out that this isn’t a driving test; nobody’s life will hang in the balance if I don’t have the knowledge I claim to possess on a certification exam.

That extends beyond just me. Having an AWS certification can definitely have benefits for folks — specifically those early in their career or transitioning to cloud after a career spent doing things that were moderately less soul-sucking.

But no hiring manager worth their salt is going to make “do you have a certification” the basis of a hire/no-hire decision. If they do, they’re wrong. Full stop.

I also point out that there’s a pandemic going on. A bit of grace to folks in these trying times does indeed go a long way.

I had to take the test on a couch in my video studio (pro tip: that thing SUCKS for sitting upright taking a test for an hour) because nowhere else in my home would pass muster. This place is open and airy, which means that “other people might be heard” which would invalidate the test.

Those other people live here; Pearson does not. I had to take pictures of my photo ID, the entire room, and then demonstrate to the proctor’s satisfaction that there was nothing on or within reach within the room.

It didn’t thrill me at all. And it didn’t reflect the reality of any business experience I’ve had at any point in my career.

What I’d change

If I could create my own AWS certification, it would focus on being a lab-driven environment where the initial setup was left to me.

Once that was complete, I’d then have the test-taker have to modify that environment while it was under load—and keep the environment up the entire time.

While this would be more work for the testing organization, I find that I don’t really care.

The point is to use the certifications as a stamp that says “I know what I’m doing.” If it’s harder to get people across the line that way? Maybe the certification was never testing the things that really mattered in the workplace.

Sure, your enterprise needs to certify the 5,000 sysadmins it’s been employing in a pre-cloud time; I get it. It takes work to teach people to cloud effectively, and more work to measure that they can do it.

Certifications have always been a trailing function. They measure competence at a point in time of an AWS service, and fail to account for the ongoing evolution of the platform.

The more of these certifications I take, the less weight I give them when evaluating where someone is with the platform. Seems a bit counterproductive, no?