Jun 16, 2026
Guest:
Kubernetes may be mature, but that does not mean its hardest problems are behind it.
Davanum 'Dims' Srinivas talks through what comes after early Kubernetes adoption: how the project keeps evolving, why maintainers are under growing pressure, and how sustainability depends on stronger boundaries around what belongs inside Kubernetes and what should remain outside it.
Bart Farrell: Who are you? What's your role and where do you work?
Davanum Dims Srinivas: Hi, my nickname is Dims. I work at NVIDIA. I've been A longtime contributor to Kubernetes. I have some leadership positions in the community as well.
Bart Farrell: Last night at dinner, talking to some folks about the evolution of Kubernetes, A lot of actions that have happened in the CNCF based on the OpenStack days that then transferred in the CNCF, the building of Kubernetes. Now Kubernetes is seen as being mature. What are the next steps?
Davanum Dims Srinivas: A funny story. When I joined the Kubernetes community, I was the OpenStack person. No, I'm not. So that's a really funny thing that happened to me. However, we see a lot of people from the OpenStack community here also. But that doesn't mean it's only the OpenStack community. We have a large number of people from various fields and from various companies and from various backgrounds that are here in the Kubernetes too. What's next? We have to figure out what to do next. We are constantly coming up with challenges and responses to challenges. This example that I was giving today in the keynote was around the AI workloads and how we could accommodate AI workloads in Kubernetes. We don't know what's going to come up next. The agent stuff is very interesting these days. There might be some more work that we'll have to do to make sure that agent workloads work well on Kubernetes too. Who knows? It's constantly evolving. The ecosystem is constantly evolving and our responses have to constantly evolve and keep pace with the technology that is evolving as well.
Bart Farrell: Very good. You've done just about every role possible inside the CNCF. I feel like I've just seen your name in So many different places in so many different roles, participating in many different ways. What have been your biggest learnings? What have been your biggest challenges?
Davanum Dims Srinivas: The biggest learning is how resilient the community is, how unique everybody's use cases are. And I can never imagine what people do with my code. Every time you talk to somebody, you come up with something new they do. I was just talking to somebody and they were talking about their challenges around deploying Kubernetes and maintaining Kubernetes and upgrading Kubernetes. There are some common themes, but everybody's position is unique and everybody's story is unique. So that is what helps us a lot because that is the feedback that we need to make it. easier and better for us to make sure that we don't break you and we evolve with the times. In terms of what has been problematic, there is a lot of burnout in the community as well. So we have to consider the responsibilities that we put on people and we have to make sure that they are supported and things like that. So there is that aspect. And with the AI stuff, it's becoming very easy for people to create PRs or CVE reports, GHSA advisories and things like that. And I'll give you one specific example. So if you go look at one of the AI-based reports that are coming in, and you will see the same person has opened 20 GHSA or PRs or issues across 15 different repositories, and they're just using AI tools to do whatever they are doing. But then as a maintainer, you are expected to actually read the one that you got and you have to answer it correctly. So it is very asymmetric between the people who are spraying the things and the people who are on the other side that have to deal with the problems and answer the questions and make sure that it is not an actual real vulnerability that can be exploited and so on. So that increases the friction, increases the angst and the worries of the people who are taking care of projects. So we have to figure out a better way to support the whole ecosystem, the set of maintainers that are taking care of the projects on their own time, essentially.
Bart Farrell: When we're thinking about Kubernetes turning 12 years old in June, and some recent developments around deprecation of Nginx ingress, does it seem like in the future people can expect more deprecations as maybe some technologies fade out, or there's a clear winner.
Davanum Dims Srinivas: The way to look at it is for Kubernetes to work for you, Kubernetes needs to be sustainable. And the way to do it is draw strong, hard boundaries around what is in, what is out, what is in Kubernetes itself, what is around Kubernetes. You go to the CNCF landscape, you can see so many projects that are built on top of Kubernetes. So that is a good sign. Earlier, we used to have a Kubernetes incubator. So that's a whole longer story. But the point is, we are at a point where when we get a set of incoming requests that are of similar nature, the Kubernetes SIGs and the SIG architecture figures out hey, maybe it's time to build an extension point. So that is a model that we are going after. Yes, Ingress Nginx is not really part of the Kubernetes binaries we ship with the release itself. So yes, it was good while it lasted. It was awesome. Everybody loved it But there is a cost associated. We were talking about maintainer burnout and mostly there was also a problem with like the CVE surfaces and security related aspects of the NGINX ingress too. I think that was more than what the existing maintainers could have handled. And maybe it was the right time to deprecate it because there are So many alternatives now. so many better alternatives. So maybe it's time to push ourselves to say, hey, we should switch to the newer alternatives rather than just sticking with the old things that we already know and love.
