This page explains the quotas and limits for Google Kubernetes Engine clusters, nodes, and GKE API requests.
GKE's per-project limits are:
- Maximum of 50 clusters per zone, plus 50 regional clusters per region.
GKE's per-cluster limits are:
- Maximum of 5000 nodes per cluster.
- Maximum of 1000 nodes per node pool.
- Maximum of 1000 nodes per cluster, if you use the GKE ingress controller.
- Maximum of 110 Pods per node.
- Maximum of 300,000 containers per cluster.
The rate limit for the GKE API is 10 requests per second.
In addition to per-cluster and per-project limits, specific GKE features have additional limits. For more information, refer to documentation for the features you use.
You might also encounter Compute Engine resource quotas. Additionally, for projects with default regional Compute Engine CPUs quota, container clusters are limited to three per region.
Resource quotas
Starting with GKE 1.12.5, a set of resource quotas is
automatically applied to clusters with five nodes or fewer and to Namespaces on
those clusters. These quotas, which cannot be removed, protect the cluster's
control plane from instability caused by potential bugs in applications deployed
to the cluster. To examine these quotas, use the
kubectl get resourcequota gke-resource-quotas -o yaml command. To view the
gke-resource-quotas object for a given Namespace, specify the Namespace by
adding the --namespace option.
Starting August 26, 2019, a set of resource quotas is automatically applied to
clusters with ten nodes or fewer and to Namespaces on those clusters. These
quotas, which cannot be removed, protect the cluster's control plane from
instability caused by potential bugs in applications deployed to the cluster. To
examine these quotas, use the kubectl get resourcequota gke-resource-quotas -o yaml
command. To view the gke-resource-quotas object for a given Namespace, specify
the Namespace by adding the --namespace option.