Nothing is ever as simple as it seems. Starting off is easy, but as your clusters grow in size and number, things will get more complex. Overall though, the “k8s is too hard” camp is just fear mongering.
Nothing is ever as simple as it seems. Starting off is easy, but as your clusters grow in size and number, things will get more complex. Overall though, the “k8s is too hard” camp is just fear mongering.
I doubt I’ll go to a NA Kubecon as long as they are in the US.
IT’S NOT A PHASE!
Awesome! Does this come in a countertop version? 🤣
FluxCD + Kustomize gets the job done
Usually, I’ll diff from what’s running with kubectl.
Too much jsonnet for me.
Will these talks be recorded?
Pulumi can be written in typescript, def worth checking out.
Such a good tool when working with #cilium.
I’d be weary of setting limits, at least for cpu. Run load tests on your apps, find where the theoretical limits are, then set your request to that.
Things rarely break for no reason. K8s is no different. Computers only do what we tell them.
Pulumi is a lot of fun. I’m thinking I should look into it again, it’s been a few years and it loos like they have made a lot of improvements.
Just because you can, doesn’t mean you should.
I’m all for k8s, but that would just be a waste of time and energy.
SOPS is what I’ve been using. It’s fairly lightweight and works well.
I’ve been running this for a bit on my clusters. It’s super handy to browse the clusters. The map feature is pretty cool too.
Have a look into k3s. It’s a single binary that lets your run k8s. Kind is also good, it stands for k8s in docker. It will set up a kubernetes cluster in docker for you.
Oops!
It’s a big onion.
There’s always some corner case . X app does 95% of what we need, but I guess we’ll have to write our own to get get to that last 5%. Then, a year later, we’re only 50% there. Ugh..
Anyone else have to deal with teams that suffer from NIH en.m.wikipedia.org/wiki/Not_inv... syndrome? #devops #k8s #kubernetes
I’d look into #cilium. Among its other advantages, it lets you assign non-routable up ranges to your cluster. So, you can have all of your pods, services and endpoints on say. 10.0.0.0/8 or 192.168.0.0/16 . Then your cluster takes up much less up space, maybe 4 IP per node.
Nice write up!
Priorities!! K8s, then poop in a fascist shoe.
You lost me at Oracle, gonna have to pass on that one.
#protip ^
Hah!! Got it! I’ll brush up on my COLBOL jic! 😂