Why does GKE Autopilot not have a pricing page?

Hi,

 Very curious to understand why GKE Autopilot is not included in the Google Pricing page. I hope Google does a better job with pricing calculators if possible.  I understand to some extent the complex math is because of the nature of the service but sometimes it is very hard to estimate costs to our customers. 

Also I am looking at the Autopilot pricing page and i don't understand a few things

1. For the  "GKE Autopilot Pod Memory Price (GB)" column let's say my POD is using 256MB so will it pro rated or will I be charged for 1GB?

2. What does ephemeral storage refer to? Nowhere in the k8s deployment manifest do i specify storage so I am not sure what this is referring to.

Solved Solved
3 3 134
1 ACCEPTED SOLUTION

1) Pricing calculator:  There are still several SKUs missing from the latest calculator, but they are being worked on.
2) You will only be charged for the memory requested by your pod.  There are certain CPU:Memory ratios that are required and Autopilot does adjust requested resources to match the required ratio.  See https://cloud.google.com/kubernetes-engine/docs/concepts/autopilot-resource-requests#compute-class-m...

3) All pods have ephemeral storage ... it's the local file system for any writes within the pod.  If you don't specify a value for ephemeral storage under resource requests in your pod spec, Autopilot sets a default (currently 1GB)

View solution in original post

3 REPLIES 3

1) Pricing calculator:  There are still several SKUs missing from the latest calculator, but they are being worked on.
2) You will only be charged for the memory requested by your pod.  There are certain CPU:Memory ratios that are required and Autopilot does adjust requested resources to match the required ratio.  See https://cloud.google.com/kubernetes-engine/docs/concepts/autopilot-resource-requests#compute-class-m...

3) All pods have ephemeral storage ... it's the local file system for any writes within the pod.  If you don't specify a value for ephemeral storage under resource requests in your pod spec, Autopilot sets a default (currently 1GB)

Thank you @garisingh Sir.

@garisingh Sorry I have 1 final question. Let's say i have a list of microservices with resource requests (CPU and Memory) that is documented and I would like to estimate the cost for GKE Autopilot in us-central1 region . Can i use the table mentioned below? For example for  a microservice with 1.25 vCPU and 285MB should i be prorating based on the table below assuming it is referring to 1vCPU and 1GB memory?

Screenshot 2024-05-07 at 5.16.30 PM.png

Top Labels in this Space