Managed Kubernetes

For most organizations, managed Kubernetes is a better choice than self-managed Kubernetes. Running self-managed Kubernetes clusters means you have to do everything yourself:

• Deploy the underlying infrastructure
• Manage the control plane
• Provision and manage nodes
• Ensure security and resiliency
• Have a team of highly skilled engineers
• Deal with a large operational overhead as a result of all of the above
  
  A managed Kubernetes provider takes care of all these aspects. However, if you want more control over your Kubernetes clusters, self-managed Kubernetes may be right for you. Your choice of how to run Kubernetes depends on your organization’s specific needs and requirements.

There are four key advantages to running a Kubernetes cluster on Bare Metal:

Performance. Clusters running on Bare Metal have no hypervisor overhead, so they can provide better performance than clusters running on virtual machines. This is especially important for applications that require low latency and high throughput.

Control. Bare Metal gives you complete control of the underlying infrastructure. This means you can choose the hardware that best suits your needs and configure it accordingly.

Latency. When you run a Kubernetes cluster on Bare Metal, you don’t have an extra virtual networking layer. Applications communicate directly over physical network interfaces. As a result, you get lower network latency.

Security. Utilizing Kubernetes on Bare Metal offers more protection than VMs because there is no shared infrastructure with other users. You have full control over security measures and can apply customized security protocols.

Read more about this topic in our article.

No, you can only manage worker nodes. Gcore takes care of the control plane management. One of the benefits of having a managed Kubernetes service is a reduced operational overhead, because you don't have to support master nodes yourself. This also ensures the reliability and security of your Kubernetes clusters.

Yes, you can. To learn more about installing nginx Ingress Controller, see our documentation and read the tutorial on installing nginx using Gcore Managed Kubernetes and Helm.

Yes, we do. You can select GPU instances to run your worker nodes during the cluster creation.

We provide a 99.9% SLA. This means that Gcore Managed Kubernetes is suitable for running your production clusters with applications that require high availability.

Gcore Managed Kubernetes supports Ubuntu for VM and Bare Metal worker nodes.

There are no Gcore restrictions for applications and tools already running in Kubernetes, including open-source options. However, there may be restrictions for tools that are managed services that only work with a specific vendor. For example, if you have been using another managed Kubernetes service along with built-in managed logging to collect logs from your clusters, that managed logging may not work with your clusters running in Gcore Managed Kubernetes. Check which of your existing tools have potential limitations for integrating with other managed Kubernetes services.

No. You are only charged for instances ordered as worker nodes. You can also buy additional services and options, such as a floating IP.

We don’t charge for egress traffic when you use VM worker nodes. However, egress traffic using Bare Metal worker nodes is charged.

Suppose you run your cluster on two VM worker nodes with a basic configuration of 2 vCPU, 4 GB RAM, 25 GB SSD, and 4 TB of egress traffic. Such a worker node costs $37.41 per month with traffic included and no control plane management fee. So your cluster will cost you $74.82 per month. You can estimate your cost using our Cloud Calculator.

You will be charged for the number of nodes you use hourly when autoscaling is in effect. You can configure the limits beyond which your cluster won’t scale.