Managed Kubernetes node configurations
In a Managed Kubernetes cluster, configurations on a cloud server and configurations on a dedicated server can be used to create node groups.
To see the availability of configurations in the regions, see the Managed Kubernetes availability matrix.
When selecting a configuration, keep in mind that some resources are reserved for system components, so the number of resources available to host pods differs from the number specified in the configuration.
Node configurations on the cloud server
Three types of configurations are available for nodes on the cloud server:
- arbitrary configurations — any resource ratio can be specified;
- fixed configurations with GPUs — ready configurations of nodes with GPUs and with specified resource ratio;
- other fixed configurations (flavors) of cloud servers — Several lineups with different specifications that have fixed resource ratios. Can only be added via API Managed Kubernetes or Terraform.
You can compare the available configurations in the Table Comparing Node Configurations on a Cloud Server.
If the available configurations are not suitable for you, for example, you need more vCPU or RAM, create a ticket — we will find configurations with other amount of resources.
Arbitrary configurations
In arbitrary node configurations, you can specify the required number of resources. Limits depend on the pool, you can see them in the table Comparing Node Configurations on Cloud Server.
You can select a local disk or one of four types of network disks as the boot disk for node.
Fixed configurations with GPU
You can choose a fixed configuration with dedicated GPUs and a specified resource ratio. For detailed GPU specifications, see the Create a Managed Kubernetes Cluster with GPUs instructions.
The configurations are in line with the GPU Line line of cloud servers.
You can select one of the four types of network disks or a local disk as the boot disk for node. For a network disk you can select the disk size, for a local disk the size is fixed.
Comparison of node configurations on a cloud server
If standard Managed Kubernetes configurations are not suitable, fixed cloud server configurations can be used.
Managed Kubernetes configurations
Fixed cloud server configurations
* If the configuration has more than 8 vCPUs, you can select RAM with a ratio of at least 1:2. For example, if you select 10 vCPUs, the RAM must be at least 20 GB.
** If there are more than 8 vCPUs in the configuration, you can select a local disk size with a ratio of at least 1:32. For example, for 10 vCPUs, the disk size is at least 320 GB.
You can see the availability of configurations in the Managed Kubernetes and GPU availability matrix for cloud servers and Managed Kubernetes.
To see the availability of fixed cloud server configurations, see the Cloud Servers availability matrix.
Node configurations on a dedicated server
In Managed Kubernetes clusters, you can create nodes on dedicated servers of off-the-shelf configurations that meet the requirements:
- there's a local port;
- x86 processor;
- 1 Gbps network card;
- without a GPU.
View detailed information about configurations in the configurator.
Dedicated servers are assembled and rack-mounted in advance, ready to work within one hour after the creation of a cluster or a group of nodes (excluding the time for OS installation).