I recently passed the Certified Kubernetes Administrator (CKA) exam. It was a great experience, and honestly the first time I’ve ever taken a tech certification and had fun with it. The CKA is a little different than your average technical certification, being a practical exam in which you are allowed to use the official Kubernetes.io documentation to formulate your answers. This is a great way to test, in my opinion, because it actually gauges your ability to use Kubernetes (or k8s, for short) in day-to-day operation – as opposed to being a glorified multiple-choice vocabulary exam.
I did not think the exam was overly difficult, and I don’t think it has to be overly difficult for you either. Just follow these five tips, and you’ll be passing your CKA in no time:
1. Study well, and play with the platform.
This sounds like two separate pieces of advice, but they are closely linked. As mentioned, the exam is a practical – it takes place in a managed lab environment in which you are directly interfacing with the k8s API. So in order to feel confident while taking the test, it’s important to not only know the concepts of k8s but how to use it as well.
To prepare, I studied the Certified Kubernetes Administrator course on ACloudGuru. I did not have to pay for it in my case, as my employer has a subscription – but you can buy it separately for your own use if necessary. This course truly did teach me the basics of k8s operation and not simply how to pass the test – so I consider it well worth your time. The major value is in the lab exercises, where you’ll write k8s manifest files and see the platform in action. With a combination of understanding the high-level concepts and understanding the actual rubber-meeting-the-road of API use, you’ll be well-equipped to pass the exam on your first try.
2. Familiarize yourself with the documentation.
The CKA is an open-book exam and you’re allowed to use any of the official documentation found on Kubernetes.io during the exam (but NOT the k8s forums, mind you). This is awesome because it reflects how an engineer would work in real life – nobody but the most eccentric geniuses actually memorize the documentation of a given system. The rest of us are constantly hopping into documentation during day-to-day work.
I relied on k8s documentation to get me through the hands-on exercises in the ACloudGuru course and I quickly became very familiar with it. This ended up serving me well during the test because I immediately knew, or at least had some idea, of where to find the information I needed to complete a given challenge. “Deploy a pod with logging sidecar? Ah, I know where to find this info.” I think this saved me valuable time on the test and helped me find what I needed more quickly – which, of course, meant more time to spend on the true brain-ticklers.
3. Verify your work.
One nice aspect of the practical exam is that you have an opportunity to actually validate the correctitude of your answer to some degree. K8s is not a system which will accept nonsense; if you provide a manifest file with significant errors – typos, invalid architecture, etc – it will usually tell you, often as soon as you try to save the file. In addition, you’re expected to actually create infrastructure, not just build the manifest – so you need to be able to run successful “kubectl create” commands. If you try to deploy a manifest and it fails, then you know you need to get back to the drawing board. This is in contrast to your average multiple choice exam in which you just have to pick an answer and hope for the best with no real clues as to whether or not you may have been right.
So in short, just pay attention to what happens when you try to save a manifest or create infrastructure. This will help you know if you’re doing things correctly or not.
4. Don’t spend more than 5 minutes on a question your first time around.
The exam is 2 hours in length with 17 questions – but don’t be fooled by that high time-to-question ratio; you’ll need every minute. Doing the math, you theoretically have 7 minutes per question. Some of the questions were VERY easy “gimmes”, especially if you’ve followed tips #1 and #2 already – so this allows you to dedicate a little more time to the difficult questions if you push yourself to just get through every question first.
If you can’t complete a challenge in 5 minutes during your first pass through the questions, then set that challenge aside and come back later – get through the rest of the questions first, so you know what kind of time you can dedicate at the end to the more difficult challenges.
In my testing experience, the hardest challenges seemed to be in the middle of the exam. I’m not sure if questions are randomized or not, so maybe your experience would be different – but either way, I recommend pushing through and answering what you can first, then use your remaining time on the tough questions.
5. Copy terms and words from the challenge descriptions, instead of typing manually.
The exam is graded using an automated process – meaning, there will be no human element. There is no sympathetic teacher who will give you a pass because they can tell you “know the idea”, despite having messed up the directions. The challenge will ask you to, for example, spin up a pod named “test-pod” with certain specifications. You might do everything perfect EXCEPT mispell the name (“test-podd”) – and you will get the question wrong because the automated grading process will not see the pod, as it’s looking for the correct name.
They are very clear on this during the exam tutorial – I recall recommendations in which they strongly suggested the test-taker copy/paste provided terms and names directly from the instructions so as to minimize the potential of being marked wrong simply for a typo. Copying and pasting will make this far less likely.
In addition, it saves time. Some of the names and terms and take a while to type – save yourself a few minutes overall and just copy/paste where possible rather than manually typing everything.
And surprise! One bonus tip which was more specific to me: Know how to backup and restore an Etcd database.
I don’t know my exact score on the test (you must get a score of at least 66% to pass, which is generous) and I haven’t tried to look it up – but I felt very confident for every question on-the-whole EXCEPT for a question in which you had to backup and restore an Etcd database. I was not able to complete that challenge no matter how I tried – and I see now what I was doing wrong.
The Kubernetes.io documentation was not quite enough to get me through it – there were a few key steps which are not explained in documentation. As to whether or not this means their documentation needs to be improved, I cannot say – but if you want to brush up on this process, this article will set you straight: “Backup and Restore Etcd.”
Best of luck on your CKA exam!