Kubernetes Resources Part 5: Vertical Pod Autoscaling

Today we are ending our series on Kubernetes Resources with a final part on Vertical Pod Autoscaling. We started the series looking at resources, what they were and how they were used and asked what autopilot for Kubernetes would look like? Are we there yet? How much effort will it be?

There are not that many resources on Vertical Pod Autoscaling, especially not its use in large scale production environments. But the material that is out there is great, check it out.

P.S. A little article on Helm included to spice things up.

Issue #60

Still up to date overview of the beta Vertical Pod Autoscaler and the use case for it: the burden of right sizing.

VPA starting in at 11:17 => Excellent overview and demo of the vertical pod autoscaler. Also though not necessarily about how they work together there is some nice discussion about that in the Q&A.

Included here because there is an announcement for a Vertical Pod Autoscaler for GKE Advanced. Does anyone have the inside scoop on how this is different from the beta VPA available through vanilla GKE?

A great step by step on setting up the VPA yourself, how to use the update policies ect. Plus its done for a tricky Java workload. Read more here.

Really lovely visual case study on using the VPA, though it’s only one pod, you can really see what the operator's experience of using the VPA is (an where it fails).

An awesome overview of not just Helm, but also Kustomize, Helm3, Tiller and more. TIL: Why Helm is so popular, it’s not just templating but it’s release management and chart repository enablement is 👌

@tryexcept’s post (One Month of Kubernetes VPA) is a good toy example of this but still wondering if anyone has experience using VPA at scale.