pulumi_eks NodeGroupV2 auto_scaling_group_tags not follow tag "Name"

[Careroms]

What happened?

From pulumi_eks==4.0.1

This is configuration
EKS_WORKER_DATASTORE = SelfManagedNodeGroupConfig(
name=f"{ENVIRONMENT}-eks-self-managed-worker-datastore-1",
cluster=eks_control_plane,
instance_profile=eks_worker_profile,
instance_type="t3.xlarge",
min_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
max_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
desired_capacity=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
node_security_group=eks_worker_sg,
cluster_ingress_rule=eks_control_plane.eks_cluster_ingress_rule,
node_subnet_ids=[
vpc.private_subnet_ids[0]
],
node_associate_public_ip_address=True,
enable_detailed_monitoring=True,
kubelet_extra_args="--max-pods=110",
labels={
"node-role/datastore": "true",
},
auto_scaling_group_tags={
"Name": "datastore",
"Purpose": "application",
"managed-by": "Pulumi",
"Environment": ENVIRONMENT,
},
taints={
"node-role/datastore": TaintArgs(
effect="NoSchedule",
value=True
)
},
)

Passing parameter to class
class SelfManagedNodeGroupConfig(CustomBaseModel):
name: str
cluster: eks.Cluster
instance_profile: InstanceProfile
instance_type: str
min_size: int
max_size: int
desired_capacity: int
node_security_group: aws.ec2.SecurityGroup
cluster_ingress_rule: Union[Output[aws.ec2.SecurityGroupRule], aws.ec2.SecurityGroupRule]
node_subnet_ids: List[Union[str, Output[str]]]
node_associate_public_ip_address: bool = False
enable_detailed_monitoring: bool = False
kubelet_extra_args: Optional[str] = None
labels: Optional[Dict[str, str]] = None
taints: Optional[Dict[str, TaintArgs]] = None
auto_scaling_group_tags: Optional[Dict[str, Union[str, Output[str]]]] = None
launch_template_tag_specifications: Optional[List[LaunchTemplateTagSpecification]] = None

And create with function

def create_self_managed_node_group(node_group_config: SelfManagedNodeGroupConfig):
node_group = eks.NodeGroupV2(
resource_name=node_group_config.name,
cluster=node_group_config.cluster,
instance_profile=node_group_config.instance_profile,
instance_type=node_group_config.instance_type,
min_size=node_group_config.min_size,
max_size=node_group_config.max_size,
desired_capacity=node_group_config.desired_capacity,
node_security_group=node_group_config.node_security_group,
cluster_ingress_rule=node_group_config.cluster_ingress_rule,
node_subnet_ids=node_group_config.node_subnet_ids,
node_associate_public_ip_address=node_group_config.node_associate_public_ip_address,
enable_detailed_monitoring=node_group_config.enable_detailed_monitoring,
kubelet_extra_args=node_group_config.kubelet_extra_args,
labels=node_group_config.labels,
auto_scaling_group_tags=node_group_config.auto_scaling_group_tags,
taints=node_group_config.taints,
launch_template_tag_specifications=node_group_config.launch_template_tag_specifications
)
This is what is created

My expectation is Name should follow configuration, if not have, default is ok.

Example

Short explanation.

1. I create auto_scaleing_group_tags as config and passing it to pulumi_eks NodeGroupV2
2. The other tags work as expected only name is replaced by cluster name + "worker"

Output of pulumi about

Version 3.187.0
Go Version go1.24.5
Go Compiler gc

Plugins
KIND NAME VERSION
resource argocd 1.0.1
resource aws 7.2.0
resource awsx 3.0.0
resource command 1.1.0
resource docker 4.8.0
resource docker-build 0.0.12
resource eks 4.0.1
resource gitlab 9.2.0
resource kubernetes 4.23.0
language python 3.187.0

Host
OS ubuntu
Version 24.04
Arch x86_64

Dependencies:
NAME VERSION
pipreqs 0.5.0
pulumi_argocd 1.0.1
pulumi_awsx 3.0.0
pulumi_command 1.1.0
pulumi_eks 4.0.1
pulumi_gitlab 9.2.0
pydantic 2.11.7
tinycss2 1.4.0

Additional context

No response

Contributing

Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[guineveresaenger]

Hi @Careroms - thank you for filing this issue.

To help us better understand the details of what you're seeing, could you please:

1. reformat your comment so the code snippets present as valid Python (with indents etc)
2. tell us a bit more about what you're expecting to see vs what you're seeing
3. When you say, "My expectation is Name should follow configuration, if not have, default is ok." what do you mean by configuration, and where are you setting this Name tag? Seeing your program formatted would help here.

Thank you!

[Careroms]

@guineveresaenger
Sorry for inconvenience I will update data as your required.

first is my custom class configuration

EKS_WORKER_DATASTORE = SelfManagedNodeGroupConfig(
  name=f"{ENVIRONMENT}-eks-self-managed-worker-datastore-1",
  cluster=eks_control_plane,
  instance_profile=eks_worker_profile,
  instance_type="t3.xlarge",
  min_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  max_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  desired_capacity=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  node_security_group=eks_worker_sg,
  cluster_ingress_rule=eks_control_plane.eks_cluster_ingress_rule,
  node_subnet_ids=[
    vpc.private_subnet_ids[0]
  ],
  node_associate_public_ip_address=False,
  enable_detailed_monitoring=True,
  node_root_volume_size=100,
  node_root_volume_type="gp3",
  kubelet_extra_args="--max-pods=110",
  labels={
    "node-role/datastore": "true",
  },
  auto_scaling_group_tags={
    "Name": "datastore",
    "Purpose": "application",
    "managed-by": "Pulumi",
    "Environment": ENVIRONMENT,
  },
  taints={
    "node-role/datastore": TaintArgs(
      effect="NoSchedule",
      value=True
    )
  }
)

Now I like to focus on auto_scaling_group_tags section

auto_scaling_group_tags={
    "Name": "datastore",
    "Purpose": "application",
    "managed-by": "Pulumi",
    "Environment": ENVIRONMENT,
  },

Next is my class detail

class SelfManagedNodeGroupConfig(CustomBaseModel):
  name: str
  cluster: eks.Cluster
  instance_profile: InstanceProfile
  instance_type: str
  min_size: int
  max_size: int
  desired_capacity: int
  node_security_group: aws.ec2.SecurityGroup
  cluster_ingress_rule: Union[Output[aws.ec2.SecurityGroupRule], aws.ec2.SecurityGroupRule]
  node_subnet_ids: List[Union[str, Output[str]]]
  node_associate_public_ip_address: bool = False
  enable_detailed_monitoring: bool = False
  kubelet_extra_args: Optional[str] = None
  labels: Optional[Dict[str, str]] = None
  taints: Optional[Dict[str, TaintArgs]] = None
  auto_scaling_group_tags: Optional[Dict[str, Union[str, Output[str]]]] = None <-<-<-
  launch_template_tag_specifications: Optional[List[LaunchTemplateTagSpecification]] = None
  node_root_volume_delete_on_termination: Optional[bool] = True
  encrypt_root_block_device: Optional[bool] = False
  node_root_volume_size: Optional[int] = None
  node_root_volume_type: Optional[str] = None

I pass my class to pulumi eks NodeGroupV2

def create_self_managed_node_group(node_group_config: SelfManagedNodeGroupConfig):
  node_group = eks.NodeGroupV2(
    resource_name=node_group_config.name,
    cluster=node_group_config.cluster,
    instance_profile=node_group_config.instance_profile,
    instance_type=node_group_config.instance_type,
    min_size=node_group_config.min_size,
    max_size=node_group_config.max_size,
    desired_capacity=node_group_config.desired_capacity,
    node_security_group=node_group_config.node_security_group,
    cluster_ingress_rule=node_group_config.cluster_ingress_rule,
    node_subnet_ids=node_group_config.node_subnet_ids,
    node_associate_public_ip_address=node_group_config.node_associate_public_ip_address,
    enable_detailed_monitoring=node_group_config.enable_detailed_monitoring,
    kubelet_extra_args=node_group_config.kubelet_extra_args,
    labels=node_group_config.labels,
    auto_scaling_group_tags=node_group_config.auto_scaling_group_tags, <-<-<-
    taints=node_group_config.taints,
    launch_template_tag_specifications=node_group_config.launch_template_tag_specifications,
    node_root_volume_delete_on_termination=node_group_config.node_root_volume_delete_on_termination,
    encrypt_root_block_device=node_group_config.encrypt_root_block_device,
    node_root_volume_size=node_group_config.node_root_volume_size,
    node_root_volume_type=node_group_config.node_root_volume_type
  )

  return node_group

The problem is when passing auto_scaling_group_tags with tag "Name": "datastore" When I look at autoscaling group that created in AWS, tag Name was replaced by staging-cluster-worker instead of datastore, but other tags is fine.
I marked my script with <-<-<- at the end of interesting line, hope it help you find them easier.

[guineveresaenger]

Thank you for clarifying! That sounds odd indeed. We'll take a look as soon as we can.

Could you help us take a look faster by providing us with a small, minimally runnable program that reproduces this behavior? It would help us see where, for example, staging-cluster-worker is coming from.

Thank you!

[Careroms]

The lower code is my eks configuration

EKS_CONTROLE_PLANE_CONFIG = EKSControlPlaneConfig(
  name=EKS_NAME,
  authentication_mode=AuthenticationMode.API_AND_CONFIG_MAP,
  skip_default_node_group=True,
  skip_default_security_groups=True,
  create_oidc_provider=True,
  version="1.31",
  vpc_id=vpc.vpc_id,
  private_subnet_ids=vpc.private_subnet_ids,
  instance_role=eks_worker_role,
  cluster_security_group=eks_ctrl_sg,
  tags={
    **SRE_TAGS
  }
)

eks_control_plane = create_eks_control_plane(EKS_CONTROLE_PLANE_CONFIG)

eks_worker_profile = create_instance_profile(InstanceProfileConfig(
  name=f"{ENVIRONMENT}-eks-worker-instance-profile",
  role_name=eks_worker_role.name,
  tags={
    **SRE_TAGS
  }
))

EKS_WORKER_GROUP_X86_1_NODE_AMT = 2
EKS_WORKER_GROUP_X86_1 = SelfManagedNodeGroupConfig(
  name=f"{ENVIRONMENT}-eks-self-managed-worker-x86-1",
  cluster=eks_control_plane,
  instance_profile=eks_worker_profile,
  instance_type="t3.xlarge",
  min_size=EKS_WORKER_GROUP_X86_1_NODE_AMT,
  max_size=EKS_WORKER_GROUP_X86_1_NODE_AMT,
  desired_capacity=EKS_WORKER_GROUP_X86_1_NODE_AMT,
  node_security_group=eks_worker_sg,
  cluster_ingress_rule=eks_control_plane.eks_cluster_ingress_rule,
  node_subnet_ids=[
    vpc.private_subnet_ids[0]
  ],
  node_associate_public_ip_address=False,
  enable_detailed_monitoring=True,
  kubelet_extra_args="--max-pods=110",
  labels={
    "node-role/application": "true"
  },
  auto_scaling_group_tags={
    "Name": "worker-group-x86-1",
    "Purpose": "application",
    "managed-by": "Pulumi",
    "Environment": ENVIRONMENT,
  }
)

EKS_WORKER_GROUP_ARM_1_NODE_AMT = 1
EKS_WORKER_GROUP_ARM_1 = SelfManagedNodeGroupConfig(
  name=f"{ENVIRONMENT}-eks-self-managed-worker-arm-1",
  cluster=eks_control_plane,
  instance_profile=eks_worker_profile,
  instance_type="t4g.large",
  min_size=EKS_WORKER_GROUP_ARM_1_NODE_AMT,
  max_size=EKS_WORKER_GROUP_ARM_1_NODE_AMT,
  desired_capacity=EKS_WORKER_GROUP_ARM_1_NODE_AMT,
  node_security_group=eks_worker_sg,
  cluster_ingress_rule=eks_control_plane.eks_cluster_ingress_rule,
  node_subnet_ids=[
    vpc.private_subnet_ids[0]
  ],
  node_associate_public_ip_address=False,
  enable_detailed_monitoring=True,
  node_root_volume_size=100,
  node_root_volume_type="gp3",
  kubelet_extra_args="--max-pods=110",
  labels={
    "node-role/application-arm": "true"
  },
  auto_scaling_group_tags={
    "Name": "worker-group-arm-1",
    "Purpose": "application",
    "managed-by": "Pulumi",
    "Environment": ENVIRONMENT,
  },
  taints={
    "node-role/application-arm": TaintArgs(
      effect="NoSchedule",
      value=True
    )
  },
)

EKS_WORKER_GROUP_DATASTORE_NODE_AMT = 1

EKS_WORKER_DATASTORE = SelfManagedNodeGroupConfig(
  name=f"{ENVIRONMENT}-eks-self-managed-worker-datastore-1",
  cluster=eks_control_plane,
  instance_profile=eks_worker_profile,
  instance_type="t3.xlarge",
  min_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  max_size=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  desired_capacity=EKS_WORKER_GROUP_DATASTORE_NODE_AMT,
  node_security_group=eks_worker_sg,
  cluster_ingress_rule=eks_control_plane.eks_cluster_ingress_rule,
  node_subnet_ids=[
    vpc.private_subnet_ids[0]
  ],
  node_associate_public_ip_address=False,
  enable_detailed_monitoring=True,
  node_root_volume_size=100,
  node_root_volume_type="gp3",
  kubelet_extra_args="--max-pods=110",
  labels={
    "node-role/datastore": "true",
  },
  auto_scaling_group_tags={
    "Name": "datastore",
    "Purpose": "application",
    "managed-by": "Pulumi",
    "Environment": ENVIRONMENT,
  },
  taints={
    "node-role/datastore": TaintArgs(
      effect="NoSchedule",
      value=True
    )
  }
)

eks_worker_group_x86_1 = create_self_managed_node_group(EKS_WORKER_GROUP_X86_1)
eks_worker_group_arm_1 = create_self_managed_node_group(EKS_WORKER_GROUP_ARM_1)
eks_worker_group_datastore = create_self_managed_node_group(EKS_WORKER_DATASTORE)

EKSControlPlaneConfig class

from modules.custom_basemodel import CustomBaseModel
from typing import Union, Optional, List
from pulumi import Output

import pulumi_eks as eks
import pulumi_aws as aws

AuthenticationMode = eks.AuthenticationMode

class EKSControlPlaneConfig(CustomBaseModel):
  name: str
  authentication_mode: AuthenticationMode
  vpc_id: Union[str, Output[str]]
  private_subnet_ids: Optional[Union[List[str], Output[List[str]]]] = None
  public_subnet_ids: Optional[Union[List[str], Output[List[str]]]] = None
  create_oidc_provider: bool = False
  skip_default_node_group: bool = False
  skip_default_security_groups: bool = False
  instance_role: aws.iam.Role
  cluster_security_group: aws.ec2.SecurityGroup
  version: Optional[str] = None
  tags: Optional[dict] = None

create_eks_control_plane function

from modules.aws.eks.models.control_plane_models import EKSControlPlaneConfig

import pulumi_eks as eks

def create_eks_control_plane(eks_config: EKSControlPlaneConfig):
  control_plane = eks.Cluster(
    resource_name=eks_config.name,
    name=eks_config.name,
    authentication_mode=eks_config.authentication_mode,
    vpc_id=eks_config.vpc_id,
    private_subnet_ids=eks_config.private_subnet_ids,
    public_subnet_ids=eks_config.public_subnet_ids,
    create_oidc_provider=eks_config.create_oidc_provider,
    skip_default_node_group=eks_config.skip_default_node_group,
    skip_default_security_groups=eks_config.skip_default_security_groups,
    instance_role=eks_config.instance_role,
    cluster_security_group=eks_config.cluster_security_group,
    version=eks_config.version,
    tags=eks_config.tags,
  )

  return control_plane

I think these are cover basic EKS configuration.