Kubernetes Engine cost vs Compute Engine Cost

[ojasgo]

I have 3 GKE autopilot clusters in 1 project. I was seeing prices are much higher than what I was expecting. turns out I am paying for compute as well as kubernetes. 

These are billing SKU that comes under Compute Engine Service.

These are the SKUs that comes under Kubernetes Engine Service 

What I dont understand is, 

I am paying for Autopilot Pod mCPU Requests (us-central1) SKU and Autopilot Spot Pod mCPU Requests (us-central1)  under Kubernetes Engine service.
There is also an SKU under Compute Engine Service Spot Preemptible E2 Instance Core running in Americas which I am paying for. Why is this the case?

I was under the assumption that VM cost would be covered under Kubernetes Engine service ONLY, as with Autopilot, I dont pay for compute but I only pay for amount of resources I request, Correct? 

This feels like I am getting double charged, I am sure I am not, but I am failing to understand what these costs actually mean !! How one SKU is different from other. 

Also, It would be great If there is a way to check cost of each GKE cluster rather than only project wide cost. Please let me know if there is a way to achieve this. 

[ojasgo]

@shannduin Anybody who can help me with this?

[shannduin]

Are you running performance class Pods in any of these clusters? 

[ojasgo]

Yes One of the cluster uses performance nodes. We removed the performance nodes from it and saw compute cost drop significantly. Is there a definite way to calculate these costs ?

https://www.googlecloudcommunity.com/gc/Google-Kubernetes-Engine-GKE/Understanding-GKE-Autopilot-pri...

I mentioned here how I am calculating the costs, but clearly this is not how it is done.

@shannduin 

[shannduin]

If removing the performance class Pods answered your question in this topic, maybe mark your answer as the solution and use the other post for an answer about pricing calculations