Alauda Service Mesh v2 Docs

Configuring Istio HA by using autoscaling

Enable High Availability (HA) for the Istio control plane through autoscaling so that the mesh remains operational even when an istiod pod fails.

NOTE

With autoscaling, you specify the minimum and maximum number of Istio control plane pods that may run at any time. Alauda Container Platform then scales the actual replica count between those bounds based on resource consumption — for example, CPU or memory utilization — so the control plane can absorb shifts in workload volume and traffic patterns within the mesh.

Prerequisites

  • You are logged in to the Alauda Container Platform web console as a user with the cluster-admin role, or you have an active ACP CLI (kubectl) session as a cluster administrator.
  • The Alauda Service Mesh v2 Operator is installed.
  • The Istio resource has been deployed.

TOC

Configuring via the web consoleConfiguring via the CLIVerification

Configuring via the web console

Procedure

  1. In the Alauda Container Platform web console, navigate to Administrator.

  2. Select Marketplace > OperatorHub.

  3. Search for the Alauda Service Mesh v2.

  4. Locate the Alauda Service Mesh v2 and click to select it.

  5. Click the All Instances tab.

  6. From the Filter by instance type dropdown, select Istio.

  7. Click the name of the Istio installation, for example default.

  8. Click the YAML tab.

  9. Update the Istio custom resource (CR) so it resembles the following example:

    apiVersion: sailoperator.io/v1
kind: Istio
metadata:
  name: default
spec:
  namespace: istio-system
  values:
    pilot:
      autoscaleMin: 2
      autoscaleMax: 5
      cpu:
        targetAverageUtilization: 80
      memory:
        targetAverageUtilization: 80
    1. spec.values.pilot.autoscaleMin sets the minimum number of Istio control plane replicas that are always running. To meet HA requirements, this value must be at least 2.
    2. spec.values.pilot.autoscaleMax sets the maximum number of Istio control plane replicas the autoscaler can scale up to under load.
    3. spec.values.pilot.cpu.targetAverageUtilization sets the target average CPU utilization as a percentage. When the observed CPU usage rises above this target, the Horizontal Pod Autoscaler (HPA) adds replicas.
    4. spec.values.pilot.memory.targetAverageUtilization sets the target average memory utilization as a percentage. When the observed memory usage rises above this target, the HPA adds replicas.

Configuring via the CLI

Procedure

  1. Retrieve the name of the Istio resource:

    kubectl get istio -n istio-system

    Example output

    NAME      NAMESPACE      PROFILE   REVISIONS   READY   IN USE   ACTIVE REVISION   STATUS    VERSION   AGE
default   istio-system             1           1       1        default           Healthy   v1.28.6   24m

    The Istio resource is named default.

  2. Patch the Istio custom resource (CR) to enable autoscaling for the control plane:

    kubectl patch istio default -n istio-system --type merge -p '
spec:
  values:
    pilot:
      autoscaleMin: 2
      autoscaleMax: 5
      cpu:
        targetAverageUtilization: 80
      memory:
        targetAverageUtilization: 80
'

    For field explanations of these settings, see Configuring via the web console.

Verification

Check the status of the Istio control plane pods:

kubectl get pods -n istio-system -l app=istiod

Example output

NAME                      READY   STATUS    RESTARTS   AGE
istiod-6b8d64c946-2dq2k   1/1     Running   0          28m
istiod-6b8d64c946-skg4v   1/1     Running   0          35s

Two istiod pods are running. Two pods is the minimum required for an HA Istio control plane and indicates that a basic HA setup is in place.