Help improve this page
To contribute to this user guide, choose the Edit this page on GitHub link that is located in the right pane of every page.
Amazon EKS Provisioned Control Plane
Overview
Amazon EKS Provisioned Control Plane is a feature that enables cluster administrators to select from a set of scaling tiers and designate their chosen tier for very high, predictable performance from the cluster’s control plane. This enables cluster administrators to ensure that the control plane is always provisioned with the specified capacity.
Amazon EKS offers two modes of operations for your cluster’s control plane. By default, Amazon EKS clusters use Standard mode, where the control plane automatically scales up and down based on your workload demands. Standard mode dynamically allocates sufficient control plane capacity to meet your workload needs and is the recommended solution for most use cases. However, for specialized workloads that cannot tolerate any performance variability due to control plane scaling or those requiring very high amounts of control plane capacity, you can optionally use Provisioned mode. Provisioned mode allows you to pre-allocate control plane capacity that is always ready to handle demanding workload requirements.
Note
Provisioned mode is an additional control plane operations mode alongside the default Standard mode. The introduction of Provisioned mode does not change Standard mode behavior.
With EKS Provisioned Control Plane, cluster administrators can pre-provision the desired control plane capacity ahead of time, providing predictable and high performance from the cluster’s control plane that is always available. EKS Provisioned Control Plane also enables cluster administrators to provision the same control plane capacity across environments, from staging to production and disaster recovery sites. This is important for ensuring that the control plane performance obtained across environments is consistent and predictable. Finally, EKS Provisioned Control Plane gives you access to very high levels of control plane performance, enabling the running of massively scalable AI workloads, high-performance computing, and large-scale data processing workloads on Kubernetes.
All existing and new Amazon EKS clusters operate in Standard mode by
default. For clusters requiring high, predictable
performance from the control plane, you can opt in to use the EKS
Provisioned Control Plane feature. You will be billed at the hourly rate
for the particular control plane scaling tier in addition to the
standard or extended support EKS hourly charges. For more information
about pricing, see Amazon EKS
pricing
Use cases
EKS Provisioned Control Plane is designed to address specific scenarios where high and predictable control plane performance is critical to your operations. Understanding these use cases can help you determine whether EKS Provisioned Control Plane is the right solution for your workloads.
Performance-critical workloads – For workloads that demand minimal latency and maximum performance from the Kubernetes control plane, EKS Provisioned Control Plane provides capacity that eliminates performance variability with control plane scaling.
Massively scalable workloads – If you run highly scalable workloads such as AI training and inference, high-performance computing, or large-scale data processing that require a large number of nodes running in the cluster, Provisioned Control Plane provides the necessary control plane capacity to support these demanding workloads.
Anticipated high-demand events – When you expect a sudden surge in control plane requests due to an upcoming event such as e-commerce sales or promotions, product launches, holiday shopping seasons, or major sporting or entertainment events, Provisioned Control Plane allows you to scale your control plane capacity in advance. This proactive approach ensures your control plane is ready to handle the increased load without waiting for automatic scaling to respond to demand.
Environment consistency – Provisioned Control Plane enables you to match control plane capacity and performance across staging and production environments, helping you identify potential issues early before deployment to production. By maintaining the same control plane tier across environments, you can ensure that testing results accurately reflect production behavior, reducing the risk of performance-related surprises during rollout.
Disaster recovery and business continuity – For disaster recovery scenarios, Provisioned Control Plane allows you to provision failover environments with the same level of capacity as your primary environment. This ensures minimal disruption and quick recovery during failover events, as your disaster recovery cluster will have identical control plane performance characteristics to your production cluster from the moment it’s activated.
Control Plane Scaling Tiers
EKS Provisioned Control Plane offers scaling tiers named using t-shirt sizes (XL, 2XL, 4XL). Each tier defines its capacity through three key Kubernetes attributes that determine the performance characteristics of your cluster’s control plane. Understanding these attributes helps you select the appropriate tier for your workload requirements.
API request concurrency measures the number of requests that the Kubernetes control plane’s API server can process concurrently, which is critical for high throughput workloads.
Pod scheduling rate indicates how quickly the default Kubernetes scheduler can schedule pods on nodes, measured in pods per second.
Cluster database size indicates the storage space allocated to etcd, the database that holds the cluster state/metadata.
When you provision your cluster’s control plane on a certain scaling tier using Provisioned Control Plane, EKS ensures your cluster’s control plane maintains the limits corresponding to that tier. The limits of control plane scaling tiers vary by Kubernetes version, as shown in the following tables.
EKS v1.28 and v1.29
| Provisioned Control Plane Scaling Tier | API request concurrency (seats) | Pod scheduling rate (pods/sec) | Cluster database size (GB) |
|---|---|---|---|
|
XL |
1700 |
100 |
16 |
|
2XL |
3400 |
100 |
16 |
|
4XL |
6800 |
100 |
16 |
EKS v1.30 and later
| Provisioned Control Plane Scaling Tier | API request concurrency (seats) | Pod scheduling rate (pods/sec) | Cluster database size (GB) |
|---|---|---|---|
|
XL |
1700 |
167 |
16 |
|
2XL |
3400 |
283 |
16 |
|
4XL |
6800 |
400 |
16 |
Monitoring control plane scaling tier utilization
Amazon EKS provides several metrics to help you monitor your control plane’s tier utilization. These metrics are published as Amazon CloudWatch metrics and are accessible through the CloudWatch and EKS console. Additionally, these metrics are scrapable from your EKS cluster’s Prometheus endpoint (see here).
| Prometheus Metric | CloudWatch Metric | |
|---|---|---|
|
API request concurrency |
apiserver_flowcontrol_current_executing_seats |
apiserver_flowcontrol_current_executing_seats |
|
Pod scheduling rate |
scheduler_schedule_attempts_total |
scheduler_schedule_attempts_total, scheduler_schedule_attempts_SCHEDULED, scheduler_schedule_attempts_UNSCHEDULABLE |
|
Cluster database size |
apiserver_storage_size_bytes |
apiserver_storage_size_bytes |
Understanding Tier capacity versus actual performance
When you select a Provisioned Control Plane scaling tier, the tier attributes represent the underlying configurations that Amazon EKS applies to your control plane. However, the actual performance you achieve depends on your specific workload patterns, configurations, and adherence to Kubernetes best practices. For example, while a 4XL tier configures API Priority and Fairness (APF) with 6,800 concurrent request seats, the actual request throughput you obtain from the control plane depends on the types of operations being performed. For example, Kubernetes penalizes list requests more than get, and hence the effective number of list requests processed concurrently by control plane is lower than get requests (see here). Similarly, although the default scheduler QPS is set to 400 for a 4XL tier, your actual pod scheduling rate depends on factors like nodes being ready and health for scheduling. To achieve optimal performance, ensure your applications follow Kubernetes best practices (see here) and are properly configured for your workload characteristics.
Considerations
-
Standard control plane capacity – EKS Standard control plane mode offers the best price to performance ratio, and is the recommended option for the vast majority of use cases. However, for specialized workloads that cannot tolerate any performance variability due to control plane scaling or those requiring very high amounts of control plane capacity, you can optionally consider using Provisioned mode.
-
Opt-in required – Existing clusters will not automatically scale up from the Standard control plane to a higher priced
EKS Provisioned Control Plane tier. You must explicitly opt in to one of the new EKS Provisioned Control Plane scaling tiers. -
Exit restriction – Standard control plane mode supports up to 8 GB of cluster database (etcd) size. If your cluster’s database size exceeds 8 GB while using Provisioned mode, you cannot switch back to Standard mode until you reduce the database size to below 8 GB. For example, if you are using 14 GB of database storage in Provisioned mode, you must first reduce your database utilization to less than 8GB before returning to Standard mode.
-
No automatic tier scaling – EKS Provisioned Control Plane does not automatically scale between tiers. Once you select a scaling tier, your cluster’s control plane remains pinned to that tier, ensuring consistent and predictable performance. However, you have the flexibility to implement your own autoscaling solution by monitoring tier utilization metrics and using the EKS Provisioned Control Plane APIs to scale up or down when these metrics cross thresholds you define, giving you full control over your scaling strategy and cost optimization.
-
Viewing current tier – You can use the Amazon EKS console, Amazon Web Services CLI, or API to view the current control plane scaling tier. In the CLI, you can run the
describe-clustercommand:aws eks describe-cluster --name cluster-name -
Tier transition time – You can use the Amazon EKS console, Amazon EKS APIs, or CLI to exit or move between scaling tiers. Amazon EKS has introduced a new cluster update type called
ScalingTierConfigUpdate, which you can inspect to monitor the progress of the transition. After you execute a tier change command, you can list the updates on the cluster to see a new update of typeScalingTierConfigUpdatewith statusUpdating. The status changes toSuccessfulupon completion of the update, or toFailedif an error occurs. The error field in the update indicates the reason for failure. There are no restrictions on how frequently you can switch between tiers. Changing the control plane tier takes several minutes to complete. -
Selecting optimal tier – To determine the optimal Provisioned Control Plane scaling tier for your cluster, you can perform load testing by provisioning your cluster on the highest tier (4XL). Then perform a load test to simulate peak demand on your cluster’s control plane. Observe the control plane tier utilization metrics at peak load, and use these observations as the guiding factor to select the appropriate tier for Provisioned mode.
-
Provisioned Control Plane pricing – You will be billed at the hourly rate for the Provisioned Control Plane scaling tier your cluster is on. This is in addition to the standard or extended support hourly charges. See Amazon EKS Pricing page
for details. -
Larger scaling tier – If you intend to run your cluster on scaling tier larger than 4XL, contact your Amazon Web Services account team for additional pricing information.
-
Kubernetes version and region support – EKS Provisioned Control Plane is supported in all Amazon Web Services commercial, GovCloud, and China regions. Provisioned Control Plane works on EKS v1.28 and higher.
