Anthos pricing overview
Anthos is available as an annual subscription, paid monthly. Anthos is sold in blocks of 100 vCPUs of schedulable compute capacity, independent of the underlying infrastructure.
The following table summarizes Anthos pricing:
(per 100 vCPU Block)
|Support||GCP support is available separately which includes support Anthos|
The subscription fee also includes a certain amount of "included usage" for several components that are priced independently but considered part of Anthos. Currently, for each 100 vCPU block these include:
- Stackdriver Logging — 5,000 GiB/month
- Stackdriver Monitoring — 375 MiB/month
- Stackdriver Trace — 50M Spans/month
- Traffic Director - 300 Endpoints (concurrent)
- How do subscription "blocks" work?
- Based on your needs, you can purchase Anthos subscriptions in blocks of 100 vCPUs. For example, if you are using Anthos to manage your workloads running on 175 vCPUs (either on-premises or in GCP), you would purchase two (2) Anthos 100 vCPU blocks for a total of 200 vCPUs. When your usage exceeds 200 vCPUs, you will be notified to purchase additional 100 vCPU blocks for the same price.
- How are charges billed?
- You will be billed at the end of each month for the subscription that you purchased.
- What is a vCPU?
- One vCPU is equivalent to one CPU as defined by Kubernetes and refers to one GCP core or one hyperthread on a physical core. Note: Oversubscribing via VMware can increase the number of vCPUs per physical core.
- How do vCPU overages work?
- You are entitled to use Anthos for the number of vCPUs that you have purchased. If your workload exceeds the vCPU blocks that you've purchased, we will notify you of the overage. There will be no impact to your running workloads. You will need to purchase additional Anthos vCPU block(s) for the remaining time on your subscription.
- How do you measure the number of vCPUs for Anthos billing?
- Anthos sizing is determined by the number of concurrent vCPUs under management.
- When is a vCPU considered "under management"?
- A vCPU is considered under management when it is seen as schedulable compute capacity by the Anthos control plane.
- Will I be charged for my Kubernetes masters?
- No. Similar to the existing GKE model, Anthos will only charge for vCPUs in user cluster worker nodes. This also applies for GKE On-Prem: you will not be charged for the Admin Cluster or User Cluster Masters.
- When can I change the size of my subscription?
- You can increase the number of blocks (vCPUs) in your subscription at any time — without extending the length of the current subscription term. You can only decrease the size of your subscription at the end of the subscription term.
- Can I split my Anthos vCPUs across on-premises and cloud or multiple projects?
- The Anthos subscription applies at the billing account. You can allocate your Anthos vCPUs across resources paid via that billing account in any way you choose. This includes:
- Split across multiple on-premises locations
- Split across on-premises and GCP
- Split across multiple GCP projects (as long as those projects are all funded by the same billing account)