drmorr

Hi bsky-ians! I rarely interact or do anything over here; mostly I just post links to my blog. All of my professional shit-posting happens over on Mastodon. But anyways follow along if you like! Here's my mastodon profile too if you care: hachyderm.io/@drmorr

Ok, after a break last week we're back with the 4th key autoscaling metric, namely, pod disruption! The fact that this was the 4th post in the series is entirely coincidental, but also hilarious, because 4 is now hilarious. blog.appliedcomputing.io/p/key-autosc... #devops #autoscaling #kubernetes

Good news, all our servers are ignoring the number 4 and the letters PHP. This security vulnerability episode brought to you by Sesame Street.

In two weeks I'm going to be presenting at OSACon (the Open Source Analytics Conference). My talk is titled "10 Useful Graphs you can't make with Grafana", wherein I will complain about all the visualization things that I wish were easier to do. osacon.io #osacon #devops #observability #grafana

whhhhaaaat the fuck? grace hopper is cancelled this year? Is this for real, or did claude hack anitab.org ?????

Ok this post was supposed to about "metrics" but I think it actually turned into a KEP. Does anybody want to write one for me? Please??? blog.appliedcomputing.io/p/key-autosc... #kubernetes #k8s #autoscaling #kep

This week on the blog we're continuing our discussion of key autoscaling metrics for Kubernetes. I'm looking at a Brand New Metric called "Mean Time to Stop Sucking", which is definitely a Real Metric that you Should Absolutely Be Tracking ™️. blog.appliedcomputing.io/p/key-autosc...

I don’t know how to explain to people that just because certain things appear to approximately work doesn’t mean they actually work. This was already horrifically difficult before AI and now it’s damn near impossible to get the point across when literally any trivial example passes muster

Setting CPU limit is still almost always wrong and setting memory limit=request is still almost always right.

In fairness, if it weren't for up to 99ms hard stalls being caused by limits, it'd probably be a lot less controversial.

Last week we kicked off our blog series about the five key metrics you should be monitoring for your autoscaler. This week, we dive into the first of these, "committed capacity". But wait! There's more, and it involves.... DaemonSets *wiggly scary fingers* blog.appliedcomputing.io/p/key-autosc...

Introducing a new blog series called "Dr. Morr says NO!" This one's about using unitless 'priority' values to signal intent in distributed systems. Don't do it, it's a trap! blog.appliedcomputing.io/p/dear-kuber... #kubernetes #k8s #distributedsystems #drmorrsaysno

OK I spent WAYYYYYYYY too long making the header image for my latest blog post, and I also immediately also put in a sticker order.

I've had a lot of other things going on so I haven't been able to write as much as I wanted this last month, but this post comes out swinging! I hope you enjoy reading it more than I enjoy working with Bazel. blog.appliedcomputing.io/p/bazels-ux-...

In this week's post, I talk about Karpenter, scheduling, bin-packing, and why it's all a giant can of worms. Enjoy! blog.appliedcomputing.io/p/karpenters... #kubernetes #karpenter #k8s

Last week my colleague Ian O'Gorman spent wayyyyy too much time debugging issues in our CI pipeline. That's a good 400 hours of his life he's never getting back. But, hey, at least he got to write a blog post about it! Buckle up, it's a banger! blog.appliedcomputing.io/p/the-seven-...

The last couple weeks I was on vacation in Korea and Japan, but my coworker Ian wrote a blog post about improving the KWOK behaviour in SimKube! blog.appliedcomputing.io/p/worth-the-... #kubernetes #k8s #kwok #simkube

Oops, I got mad and wrote a bonus blog post this week. Maybe ignore it if you also don't want to get mad. blog.appliedcomputing.io/p/stop-accus...

This week's blog post is about the work that my students at Harvey Mudd College did to turn SimKube into a reinforcement learning arena for AI agents! blog.appliedcomputing.io/p/simarena-a...

ACRL is turning 3, and drmorr is turning an age greater than 3!!! blog.appliedcomputing.io/p/acrl-is-tu...

Today's blog post is about all the techniques I can think of to mock Rust code: blog.appliedcomputing.io/p/all-the-wa... #rustlang #testing #software

In case you missed it, this is my SRECon recap, featuring all of my favorite talks! blog.appliedcomputing.io/p/srecon-202...

This talk features my postmortem of the worst incident I've ever dealt with in production! Excited to get to finally check out the talk.

Simulation testing before upgrades beats changelog roulette every time, though honestly most teams still just wing it and hope the canary doesn't explode.

Are you bored on a Friday afternoon? My SRECon talk, "Stop Reading Changelogs: Safer Kubernetes Upgrades with Simulation" is now up on YouTube. Go watch it! Feels very timely since Kubernetes (checks notes) 1.36 was just released! www.youtube.com/watch?v=yk_P... #srecon #kubernetes #simkube