Uploaded image for project: 'OpenShift Container Platform (OCP) Strategy'
  1. OpenShift Container Platform (OCP) Strategy
  2. OCPSTRAT-2410

AWS IPI: Enable setting the throughput flag for GP3 drives on master/worker instances during installation

XMLWordPrintable

    • Product / Portfolio Work
    • None
    • 0% To Do, 100% In Progress, 0% Done
    • False
    • Hide

      None

      Show
      None
    • False
    • XS
    • None
    • None
    • None
    • None
    • None
    • None

      Feature Overview (aka. Goal Summary)

      Enable customization of throughput settings for rootVolume drives during the installation of OpenShift in AWS to prevent production outages caused by the masters being capped at the default value of throughput.

      Goals (aka. expected user outcomes)

      Users will be able to customize the throughput settings of rootVolume drives at install time. This will enhance the existing feature of setting the root drives for the masters, which currently defaults to a capped throughput of 125MiB/s for GP3 volumes.

      Requirements (aka. Acceptance Criteria):

      • Allow users to specify the throughput for AWS GP3 `rootVolume` in the `install-config.yaml` file for both control plane and compute nodes.
      • The installer must correctly provision the AWS EBS volumes with the specified throughput.
      Deployment considerations List applicable specific needs (N/A = not applicable)
      Self-managed, managed, or both Both
      Classic (standalone cluster) Yes
      Hosted control planes Yes
      Multi node, Compact (three node), or Single node (SNO), or all All
      Connected / Restricted Network Both
      Architectures, e.g. x86_x64, ARM (aarch64), IBM Power (ppc6le), and IBM Z (s390x) All
      Operator compatibility Yes
      Backport needed (list applicable versions) N/A
      UI need (e.g. OpenShift Console, dynamic plugin, OCM) May be required
      Other (please specify) N/A

      Use Cases (Optional):

      A customer wants to install an OpenShift cluster in AWS. They need to configure the root volumes for the control plane and worker nodes to use GP3 disks with custom throughput settings at installation time to prevent potential production outages caused by I/O throttling.

      Questions to Answer (Optional):

      What are the technical challenges in implementing this feature within the installer?

      Out of Scope

      This feature will only cover throughput customization for GP3 volumes. Other volume types are out of scope.

      Background

      This feature request originates from production outages where control plane nodes were throttled at the default throughput for GP3 volumes. High workload pressure on the Kubernetes API server and etcd led to maximum throughput utilization, causing instability. The customer specifically requested this to avoid manually adjusting volumes post-installation across their many clusters.

      Customer Considerations

      Customers with I/O-intensive workloads who have experienced or are at risk of performance degradation and outages due to default EBS throughput limits will benefit significantly. This allows them to proactively configure their infrastructure to match their workload needs from the beginning.

      Documentation Considerations

      The official OpenShift documentation for AWS installations will need to be updated to include the new `throughput` parameter for `rootVolume` configuration in the `install-config.yaml`. A clear example should be provided.

      Interoperability Considerations

              linnguye.openshift Linh Nguyen
              linnguye.openshift Linh Nguyen
              None
              John Hixson
              John Hixson John Hixson
              Yunfei Jiang Yunfei Jiang
              Avani Bhatt Avani Bhatt
              Derrick Ornelas Derrick Ornelas
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated: