A recent nationwide U.S. search on LinkedIn, for example, turned up nearly 26,000 open positions. They include titles like “Kubernetes platform lead” and “Kubernetes administrator,” but also a wide range of other common titles: DevOps engineer, SRE, senior manager of software engineering, and many more.
This is all to say: IT leaders and recruiters are looking for Kubernetes skills across a significant range of technical roles and titles. No matter the role you’re hiring for, you’ll need to ensure you’re asking the right kinds of questions during the interview process.
[ Kubernetes 101: An introduction to containers, Kubernetes and OpenShift: Watch the on-demand Kubernetes 101 webinar.]
Kubernetes interviews: Experience is not a given
You could say that’s true of any technical skill – or just about any other type of role outside of IT – and you’d generally be correct. But the interview process will be particularly important when looking for Kubernetes skills because the technology is still relatively young.
“We love to see candidates with Kubernetes experience,” says Tom Manville, engineering lead at Kasten. “When candidates do have Kubernetes experience, it’s important to gauge their level of expertise alongside their problem-solving skills and development capabilities.”
But finding experienced candidates is not always a given, so you may need to cast a wider net and consider people with an interest and aptitude for learning on the job.
“Recognizing that this space is new and developing fast, we also look for candidates who have overcome complex problems in the past and can clearly explain their design choices,” Manville says.
[ Check out our related post: Kubernetes: 3 ways to get started. ]
In either scenario, Manville recommends a couple of starter questions to get going. They kick off our list of Kubernetes interview questions from Manville and other experts, which we’ve grouped into three categories. We start with the basics, which can be especially useful if you’re looking for someone who has taught themselves the fundamentals but is still a relative beginner.
You can, of course, mix up your preferred order or cherry-pick and adapt as you see fit. We’re approaching this from the POV of a hiring manager, but job seekers should use this to prepare, as well.
We’ve also shared some thoughts on tailoring questions toward particular roles that might use Kubernetes differently. Let’s dig into some questions you can use to build your own interview framework.
[ Why does Kubernetes matter to IT leaders? Learn more about Red Hat's point of view. ]
Kubernetes interview questions: Covering the basics
1. What is required to deploy a simple application, like a web server in Kubernetes? How would you go about deploying/updating your app?
This is a 101-level question, to be sure, but it’s a good one for weeding out folks who might be overstating their qualifications, especially given that tools like Minikube enable practitioners to do things like deploying simple apps.
“It is important for candidates to be able to articulate how they would walk through a simple project to demonstrate that they understand the fundamental Kubernetes components,” Manville says.
[ Kubernetes terminology, demystified: Get our Kubernetes glossary cheat sheet for IT and business leaders. ]
2. When defining an application, how would you choose between using the various Kubernetes workloads? For example, when would you use a Deployment versus a StatefulSet versus a DaemonSet?
“Workloads are a core concept in Kubernetes,” Manville says. “Even if a candidate doesn’t know all the details and differences between workloads, they should be able to explain them at a high level.”
The Kubernetes documentation on fundamental Kubernetes concepts has a lot of useful detail here.
3. What are container orchestrators and why are they required?
This one, from Raghu Kishore Vempati, director of technology, research, and innovation at Altran, gets at the fundamental question of why containers and orchestrators exist in the first place: What are they for? What are the benefits and challenges? Fumbling this question might be a sign that a candidate needs more learning and experimentation.
“We expect them to explain nomenclature of how several components of the same application can execute in a varying group of containers, the varied clustering and scaling requirements for components, and then [the] ability to have [an] automated approach to continuously deploy applications to a given environment,” Vempati says.
[ Read also: OpenShift and Kubernetes: What’s the difference? ]
Kubernetes interview questions to find learners and problem-solvers with high potential
Nathan Cooprider, senior manager, engineering at Threat Stack, is a proponent of expanding your parameters to look for less experienced pros with high potential.
“Kubernetes is such a new technology that it’s unrealistic to expect all job candidates to have a ton of hands-on experience,” Cooprider says. “The key to finding a successful job candidate in a Kubernetes-facing role is to focus on finding someone who has proven their ability to learn, adapt, and deal with unexpected problems in previous roles.”
Here are several of his recommended questions when considering folks with some experience, but who are still learning.
4. Tell me about a Kubernetes cluster you deployed. How did you test for load? What monitoring did you implement? What metrics did you track?
5. What experience do you have with containers? What tooling have you used to manage containers? What advantages and disadvantages did you see in the tooling?
6. Tell me about a time you had difficulty deploying new software you did not write. How did you resolve the difficulty? Where did you go for additional information?
The idea here is to get a feel for their understanding of “why Kubernetes?” If they can’t speak at all to containers and related technologies, for example, that’s a red flag, or at least something to dig into further. Question 5, in particular, speaks to the need for Kubernetes pros with a knack for troubleshooting and problem-solving, critical abilities beyond mastering the nuts and bolts of the platform itself.
The fact that Kubernetes is rapidly evolving requires an interest in ongoing learning, even for people with deeper hands-on experience. Ravi Lachhman, DevOps evangelist at Harness, recommends two questions that get at learning and problem-solving abilities, as well as (human) resiliency when failure occurs.
7. If you had to redo a project you just completed, would you be upset?
“The Kubernetes platform and ecosystem moves fast. New ways of thinking are constantly popping up, and our approaches to working with Kubernetes need to shift as the ecosystem evolves,” Lachhman says.
“For example, a package manager called Helm (a stalwart in the Kubernetes ecosystem that’s only a few years old) is already being chipped away at by another project called Kustomize,” he points out. “The introduction of Service Meshes has been driving architecture and shifting complexity and networking ownership around. It’s important to find out if a candidate is willing to redo a project to leverage new approaches for the good of the project. Would they throw up their hands, or embrace the shifting environment? This question helps an interviewer understand how candidates approach these issues.”
8. How do you like to learn new concepts? If you failed, would you be upset?
“Both parts of this question are important. Kubernetes ecosystems provide a huge (and growing) number of resources from which to learn new skills. Because approaches change so frequently, there are no one-size-fits-all concepts for working with Kubernetes, and sometimes developers will fail,” Lachhman notes.
“For example, the tried-and-true love of system administrators learning from play/run books is hard to pin to Kubernetes. Developers need to be comfortable with taking failure in stride, learning from sources that are not exclusive to their company, and sharing best practices with the community at large. Keep an open mind!”
What if you need candidates with more depth of experience? Let's examine some questions to gauge that: