You don’t work with the large enterprises that I work with everyday. Let’s not pretend that K8s doesn’t bring its own set of complications. Not to mention enterprises that have 500+ servers, some running Windows, some running COTS, etc.
Much of the time it is easier just to spend up a VM and install whatever you need.
Ahh government work. Can always tell because y’all don’t call them appliances. But I mean idk, I kinda do. The largest place I ops’d for had two 40 rack floors with some misc stuff here and there. Don’t remember at all how many of individual servers that meant once you accounted for KVMs, SANs, UPS, and networking but it was a lot.
I mean yeah we had a huge Hyper-V cluster for the Windows IT folks, bunch of random appliances bought over the years, but then for our software stack we ran a big-ole RHV cluster (look, we had a sweetheart deal from RH gave us everything) spent a lot of time getting k8s hooked up to everything (open stack we decided wasn’t worth the squeeze) and devs were v happy, no more tickets! All the backing infrastructure was vm based but we managed it and the devs got to care no percent.
> Ahh government work. Can always tell because y’all do
Uhh I work in the Professional Services department of a little cloud provider based out of Seattle you might have heard of…
Sure I’ll work on a project porting to EKS (managed Kubernetes) if they are already familiar with it. But rarely will I suggest introducing the complexity of it unnecessarily.
I’ve worked with governments, colleges, F500s, startups, and everything in between.
I mostly focus on green field app dev. But I know my way around “DevOps” [sic]
Ahh government work! Look I’m just poking fun at the fact that people who call them COTS are either government or sell to government, it’s not a pejorative.
Not doubting your expertise at all, I’ve worked at shops that are entirely vm based, I just think that when it comes to
making dev teams happy in an on-prem world giving them a k8s cluster has so far never not been a good decision. I can hide so many of the ugly realities of running a production service.
You have to talk the talk to show them that you’re one of them. I can go from a decent midwestern neutral accent to my natural southern drawl depending on which state I’m working with too :)
And besides, even when I worked in the real world as your bog standard Senior enterprise dev, I got to the point where I wouldn’t work at any company that had on prem architecture.
My main reason for going to the cloud was not that I didn’t want to deal with administering servers, I didn’t want to deal with server administrators.
Much of the time it is easier just to spend up a VM and install whatever you need.