Kubernetes FAQ

Which regions is Kubernetes available in?

Delhi and Chennai. Persistent volumes are available on clusters in Delhi. See Create a cluster.

Why can I not create a cluster in Delhi?

In Delhi, an account created on or after May 15, 2026 cannot create a cluster unless E2E has exempted it. The request is refused with Resource creation is currently restricted. Please try from a different location., and CPU plans show Request instead of being selectable. Launch your cluster in Chennai, or contact support to create one in Delhi. See Troubleshooting Kubernetes.

Can a cluster have more than one master node?

No. Every cluster has one master node, which runs the Kubernetes control plane. To give it more room, upgrade its plan. See Upgrade the master plan.

Can I give the master node more vCPU and RAM?

Yes. Upgrade Kubernetes Master Plan moves the master node to a bigger plan, with more vCPU and more RAM. The cluster keeps its Kubernetes version. The upgrade takes about 5 to 7 minutes of downtime and is one way. See Upgrade the master plan.

Can I upgrade a cluster’s Kubernetes version?

No. A cluster keeps the version it was created with for its whole life. To move to a newer version, create a new cluster with it, and deploy your applications there. See Create a cluster.

Can I rename a cluster?

No. A cluster’s name is fixed once it is created. You can rename a worker node in the portal; see Manage node pools.

How many node pools and worker nodes can a cluster have?

A cluster needs at least one worker node pool and can have up to 10. A pool holds 2 to 25 worker nodes, or 1 to 25 on the GPU or GDC 3rd generation family. See Manage node pools.

Can I use my own VPC range?

No. A cluster is built in a VPC with an E2E provided range, and the create form lists those VPCs. Every master and worker node gets a private IP from the cluster’s VPC, or from its subnet when you pick one. See Create a cluster.

Which ports must the security group allow?

Inbound TCP port 6443 and UDP port 8472; a cluster is created only with a group that allows both, and kubectl reaches the cluster through the 6443 rule. Security groups control traffic to the nodes’ own IPs. A LoadBalancer service on a public service IP is reachable on every port it exposes, whatever the groups allow, so it needs no rule of its own; to limit who reaches it, set its source ranges. See Manage a cluster’s security groups and Limit who can reach the service.

Can I SSH into the cluster’s nodes?

No. You work with the nodes through kubectl and the Node Pool actions: Start, Power Off, Reboot, and Delete. See Manage node pools.

How do I set labels and taints on nodes?

With kubectl, using the node’s Kubernetes Worker Name. See Manage node pools.

Can autoscaling follow memory use?

No. Elastic scaling follows CPU or a Custom metric that you send yourself. See Configure autoscaling for a node pool.

How is a Kubernetes cluster billed?

The master node and every worker node are E2E nodes, each billed on its own plan, by the minute or at a committed price; an autoscaling pool is always billed by the minute. Every persistent volume is billed by size from when it is created until it is removed from storage. Public service IPs are reserved IPs, billed from when you reserve them until you release them, and private LB IPs are free. For current prices, see E2E’s pricing page. For committed plans, see Use committed billing for a cluster.

Can I put an autoscaling pool on a committed plan?

No. An autoscaling pool is always billed by the minute. Committed pricing is for the master node and static pools. See Use committed billing for a cluster.

Can I resize a committed pool?

No. A committed pool keeps its size for the term; to add capacity, add another pool. See Manage node pools.

What happens when a committed term ends?

The master node or pool does what its end-of-term setting says: it renews the committed period, or it moves to hourly billing at the plan’s base rate. A new setting takes effect when the current committed period ends. On a prepaid account, auto-renewal needs infra credits for the whole master’s or pool’s renewal; without them, all of those nodes move to hourly billing instead, and E2E emails you either way. See Use committed billing for a cluster.

Can I put the E2E Load Balancer in front of my cluster?

No. Kubernetes master and worker nodes are not listed as backends in the Load Balancer product. Expose your application with a LoadBalancer service and the cluster’s LB IP Pool. See Expose a service with a load balancer.

Are private LB IPs charged?

No, private LB IPs are free. Public service IPs are reserved IPs, billed from when you reserve them until you release them. See Expose a service with a load balancer.

Can I create persistent volumes in Chennai?

No. Persistent volumes are available on clusters in Delhi. In Chennai, the cluster has no Persistent Volume tab. See Use persistent volumes.

Does the cluster certificate renew itself?

Yes, while the cluster reads Running on its renewal date, one year after the cluster is created and one year after each renewal. You can also renew it yourself in the five days before the date. After a renewal, download the kubeconfig again. See Renew the cluster certificate.

Can I delete a cluster that is stuck on Creating?

No. Delete is available once the cluster reads Running. If it still reads Creating after 80 minutes, contact support. See Troubleshooting Kubernetes.

Does my cluster have a fixed outbound IP?

No. Pods reach the internet from their node’s own public IP, and nodes that a resize or autoscaling adds get new addresses. See Manage node pools.

What if my kubeconfig leaks?

Contact support, which replaces the cluster’s credentials for you. See Connect to a cluster with kubectl.

What does deleting a cluster remove?

Its pools, its persistent volumes, the ones your claims created included, its private LB IPs, and the security groups E2E made for it. Its public service IPs go back to your reserved IPs, where they are billed until you release them. See Delete a cluster.

When does billing stop after I delete a cluster?

When each machine and volume is actually removed, and for public service IPs, when you release them. See Delete a cluster.

Does kubectl delete node stop the node’s charges?

No. kubectl delete node removes the node from Kubernetes only; its machine keeps running and is billed. Delete worker nodes from the Node Pool tab. See Manage node pools.

How do pods reach a DBaaS database?

Over the VPC. Every master and worker node gets a private IP from the cluster’s VPC, so attach the cluster’s VPC to the DBaaS instance, and connect from your pods to the database’s Private IPv4 on its port. If the database has IPs on its whitelist, add the cluster’s VPC range (its CIDR) too, so every node, including ones added later, can connect. See Attach a VPC to a DBaaS instance.

How do I back up a cluster?

With Velero, a backup tool that runs in the cluster, and an E2E Object Storage bucket. Velero backs up and restores namespaces, and a claim made through csi-rbd-sc comes back as a new volume with its data. See Back up a cluster with Velero.