Having to recycle your #kubernetes nodes because the #HPA isn't scaling up your #Mastodon#sidekiq so you're now lagging 15K jobs behind looks like this in the #UI:
Had fun this weekend working on a performance focussed proof of concept using Bunny in the #PHP#queue interop contracts. The first metrics are in using the #RabbitMQ cluster on my #Raspberrypi#Kubernetes home cluster. (Which isn't meant for high performance. Still pleased by these numbers.)
#AlpineLinux is generally my default goto for #kubernetes#container image bases when I'm writing stuff from scratch. I love how lightweight and simple it is, so I'll often start there until I can prove that I need something more complex.
On that note - need to move some #Kubernetes data to S3? Here's an #Alpine image I pushed today that includes #s3cmd and zip and gets built/tested/pushed weekly by our Jenkins farm at my $DAYJOB. Perfect for a sidecar container or CronJob backup, etc. Enjoy! (And if this is useful to you, thank the Alpine and s3cmd folks, because they by far did most of the work.)
I came up with a thing in #fsharp today and I am unsure about it. I don't know if it's cute, useless, or useful; I cannot see far enough ahead to understand its implications.
type ErrExn = | Err of DomainError | Exn of System.Exception
This DU unifies domain errors (that are represented in a user-written DU) and exceptions.
I'd expect it to be in a function with a signature of 'a -> Result<'b, ErrExn>.
I don't know if this gives value over separate domain errors and exceptions.
@pblasucci@chethusk@greggyb
"Let it bubble up and crash" is also more in line with #kubernetes and modern #microservices 10 factor app principles.
It's not "crash" though, it is "shut down gracefully"
For a worker #process, graceful shutdown is achieved by returning the current job to the work #queue.
Decided to sit down and try to learn #kubernetes despite the fact that I don't have any real need for it on my little home server. Still, it's been interesting and - I won't lie - a bit frustrating at times. It's very much like using a chainsaw to butter my bread for my use case, but I had a nice feeling of satisfaction when I succeeded in getting #owncast set up through it. I've had some odd issues with Docker failing to launch certain containers through containerd that I have not been able to figure out, however.
I was putting thought into maybe redoing my server setup with kubernetes but I sincerely worry that I'll run into this same containerd issue with some of my other apps.
Blog Update: I have successfully migrated email sending away from ButtonDown Email to a self-hosted Ghost instance. Nothing bad with ButtonDown - simply, Ghost has by far a better editing and publishing experience (and can be self-hosted, of course).
Speaking of newsletters, one reason to migrate was because I want to explore email communication a bit further. So, if newsletters are your thing, be my guest and add your email here: https://preslav.me/subscribe/#via-email I won't spam!
None of that would have been so easy, if it wasn for @andrasbacsai and his monstrous work on Coolify! If you have a spare server and want to turn it into a self-hosted #PaaS without #Kubernetes and all the jazz, Coolify is absolutely the best!
Energy use of #Kubernetes when it's not doing anything is off the charts. There's about 4 processes that just sit there doing 'something' even when there's zero happening on the cluster. Initialize k8s, start it up and do nothing. Energy being burnt. So if you have less than 8 cores on that box, then it's spamming the scheduler making everything else laggy.
I managed to avoid #Kubernetes for 10 years, but it’s finally caught up to me, so I hope I’m a Kubernetes god after going through all this required (by job) Kubernetes training.
you know hacking kubernetes manifests is so much more comfortable in python... is there any drive to get a yaml processor into the python standard lib?
Trying to automatically/programmatically replicate #DockerCompose stacks on the same host. E.g. I have an application that requires multiple containers, and I want to replicate the WHOLE application with its own volumes, networks, subdomain, etc. Any pointers on how to do that? #Ansible? Please don't say #Kubernetes.
This might come across sorta dumb but I like that I can delete all of the worker nodes of a #kubernetes cluster, keeping the control plane of course, and then put the worker nodes back and everything works perfectly fine again.
I'm still learning the tools in nicolaka/netshoot, but also need some #dotnet specific things.
I stumbled on the lightrun-platform/koolkits package for node, and I wish they had a #dotnet image -- anyone using anything better than the dotnet-monitor container?
First part of a new long term home project coming in. An #Ubiquiti PoE+ switch to power a small #Kubernetes cluster built using #raspberrypi nodes. Going to blog about every step once it has been completed. But it is going to be a few quarters long project doing bit by bit
So instead I went with this tower, we have the same one downstairs powering the living room #lego display. Plenty of space and nothing will block each other. Plus it comes with 4 #USB powers we will utilize for (more) lego display lights and planned #ESPHome sensors on the stairs to the home office. Just need to pick and time and date as doing this will take that entire switch down, and create a network split between two and one control plane #kubernetes nodes.
The fun part is that I've already accidentally ran with 3 nodes, while I need 4 over the past week occasionally. Tweaked my home brow #kubernetes cluster autoscaler a bit to aggressive and it took out a node while it shouldn't.