eks: managed node group allocatable pods number error

[tony-vsx]

Describe the bug

When I use managed node group multiple instance types, I got wrong allocatable pods number, t3a.xlarge allocatable pods
should be 58 not 17

Expected Behavior

t3a.xlarge allocatable pods should be 58 not 17

Current Behavior

t3a.xlarge allocatable pods got 17

Reproduction Steps

const cluster = new eks.Cluster(scope, 'eks', {
version: eks.KubernetesVersion.V1_29,
kubectlLayer: new KubectlV29Layer(scope, 'kubectl'),
clusterName,
vpc,
vpcSubnets: [{ subnetType: ec2.SubnetType.PRIVATE_WITH_EGRESS }],
endpointAccess: eks.EndpointAccess.PRIVATE,
role: iam.eksRole(scope),
outputMastersRoleArn: true,
albController: { version: eks.AlbControllerVersion.V2_6_2 },
clusterLogging: [eks.ClusterLoggingTypes.AUTHENTICATOR],
defaultCapacity: 0,
securityGroup: Vpc.eksControlPlaneSecurityGroup(scope, vpc),
})

new eks.Nodegroup(scope, ${clusterName}-default-group, {
cluster,
diskSize: 20,
desiredSize: 2,
minSize: 0,
maxSize: 10,
nodeRole,
subnets: { subnetType: ec2.SubnetType.PRIVATE_WITH_EGRESS },
instanceTypes: [
new ec2.InstanceType('t3a.xlarge'),
new ec2.InstanceType('t3a.medium'),
],
tags: {
[k8s.io/cluster-autoscaler/${clusterName}]: 'owned',
['k8s.io/cluster-autoscaler/enabled']: 'true',
Name: ${clusterName}-default-group,
},
})

Possible Solution

No response

Additional Information/Context

No response

CDK CLI Version

2.128.0

Framework Version

No response

Node.js Version

20

OS

MacOS

Language

TypeScript

Language Version

No response

Other information

No response

[msambol]

The first two screenshots show 58. The second set shows 17, which is the limit for t3a.medium as defined here. Can you try this again but for instance types only include t3a.xlarge ?

[tony-vsx]

It will be 58

[msambol]

So it's behaving as expected ? 😀

[tony-vsx]

If I only choose one, yes. But if I choose two, no.

[msambol]

Right, because you are also specifying t3a.medium as an instance type:

instanceTypes: [
   new ec2.InstanceType('t3a.xlarge'),
   new ec2.InstanceType('t3a.medium'),
],

If you want 58 allocable pods, only use t3a.xlarge. Hope that helps.

[pahud]

@tony-vsx did you mean you are having 17 allocable pods on t3a.medium which should be 58?

I think it's managed by the agent in the AMI as described in this file(thank you @msambol ) and t3a.medium is actually having 17 allocable pods.

[tony-vsx]

@pahud yes, that's what I want to express

[pahud]

I think it should be 17 according to that file. I am closing this issue as it doesn't seem to be a CDK bug. Feel free to reopen if there's any concern.

[github-actions]

⚠️COMMENT VISIBILITY WARNING⚠️

Comments on closed issues are hard for our team to see.
If you need more assistance, please either tag a team member or open a new issue that references this one.
If you wish to keep having a conversation with other community members under this issue feel free to do so.

[tony-vsx]

Sorry for the confusion. What I meant to say is that if set within the same array, t3a.xlarge should be set to 58, t3a.medium to 17, but when set together,t3a.xlarge would end up being only 17.