GKE Compute Cost Comparisons In Small Clusters

Sun, 21 May 2023 19:00:00 -0100

I’ve been running a short-term experiment across the GKE Compute options within my “Home Lab” (read: GCP-hosted Google Kubernetes Engine) with a view to optimise for a balance between ease of operation and cost efficiency. If that sounds useful to you, then by all means read on!

Let’s dive right in with the Billing data. The chart below shows the view of costs over the course of my two-month experiment for this small GKE Cluster. I’ll be stepping across this graph to detail each experiment I ran and what led to these outcomes, as well as some of the trade-offs along the way.

Squeezing GKE System Resources In Small Clusters

Sat, 28 Sep 2019 18:00:00 -0100

Photo by Davide Ragusa on Unsplash

Spoiler Alert! This blog is really about Vertical Pod Autoscaling and patching of kube-system workloads in GKE. It just might not sound like it at the start 😄 If you’re not interested in how I got there and just want to jump to the good stuff - I’ve summarised it at the bottom of the post.

So, this all started because I wanted to try out Elasticsearch. I’m not the biggest fan of GCP’s Stackdriver Logging, and I wanted to poke around with some alternatives. I have a “personal” GKE cluster that I use to run a few websites (including this blog!), and where I also experiment with these sort of things outside of work. Should be straight-forward I thought to myself. Sure, Elastic is Java - but I don’t need it to do that much, I’ll just shrink it down enough that it’ll start up ok, play around a bit, and then maybe see if I want to do more with it or sign up to a managed logging service instead.

VPA on alexos.dev

GKE Compute Cost Comparisons In Small Clusters

Squeezing GKE System Resources In Small Clusters