When is Kubernetes worth it?

Kubernetes isn’t just hard to spell, it’s hard to maintain, too.

However, it’s not by chance that it’s a very popular technology since it does come with a lot of benefits. The topic of Kubernetes even came up when we were discussing vendor lock-in.

With serverless technologies – mainly Functioning as a Service in this scenario – getting cheaper and easier to run at scale, my question is: When do you think Kubernetes makes sense? Is there a specific audience for this platform?

2 Likes

So far I can tell you that running it onprem seems to be too difficult to maintain. In my personal opinion, it’s only worth it if you are cloud based where you can leverage their tools or their Kubernetes infrastructure to run it. But honestly, I have to agree with you that it’s not easy to define when it’s worth it.

1 Like

I think we’re in that “fun” place right now where EVERYONE wants to use Kubernetes because it’s #TheThing right now, and it in a lot of cases makes environment more complex and harder to support than they were before the redesign.

Honestly, as a “not Kubernetes expert” I also find it hard to determine WHEN it makes sense, but I do have the clarity to know that it doesn’t always make sense to use it. So that’s not nothing. :laughing:

3 Likes

I look forward to the marvellous future when Kubernetes sinks beneath the surface of things application developers need to know about. I spent 3 years (:exploding_head:) with it in various forms and am happy to have already forgotten the majority of what I knew.

1 Like

I think your question @raphabot is what does it do in comparison to serverless?

And I agree not much if you think only about the functionality you need running.

The added complexity of maintaining all the moving parts involved is usually the deciding factor for moving away.

I guess this is a long winded way of saying that getting functionality to work on your cluster you built and configured on your own is incredibly rewarding; but it kind of falls under the bag of doing it once and moving on.

IMO, the concerns about vendor lock-in are valid but with since every single Cloud has an API, migrating is possible and not just a dream.

2 Likes

Saw this article today and I like it’s take on Kubernetes.

InfoWorld: You’re thinking about Kubernetes all wrong.

“It’s actually much better to let individual application teams set up their own infrastructure using Kubernetes clusters instead of procuring enterprise cluster resources from central IT,”

Wow. I’m not sure if I agree with that much. First, there is the cost. Running k8s is far from being cheap. If you think AWS’ EKS, it’s $0.10/hour, or about the as a EC2 instance that has 4 vCPUs and 8GiB or RAM. Second, there is the expertise to run it. k8s is far from being “next > next > finish” kind of solution, so to run it with defaults for anything more than just a lab environment is kind crazy to me. On top of that that, third in the list here, is security. Many organizations are struggling to put good security practices and hygiene around their k8s clusters, increasing the number of clusters by the same amount of apps can’t make it any easier.

In the old days, developers would start and stop Tomcat at will. They wouldn’t ask someone else’s permission to do it, right?

We definitely worked in different companies :slight_smile:

How did you feel about this article?

Look at me digging this old topic :slight_smile:

And the reason I’m doing it is that, part of the reason why I think Kubernetes is so complex, is that I don’t fully understand how to make sure I’m not leaving it too exposed. If that’s also a challenge that you might have, @marykay25, @bnwoods, @glb, and @atirado, I’d like to let you know that @magnologan will be doing a session on Kubernetes Security 101.

If you want to know more about it, and sign up for it, check it out here.

2 Likes