🚀Add functionality to use also existing agent pool in agentpool.yaml CR configuration

[zoli-opslogic]

Description

We are in the process of migrating our whole infrastructure into HCP Terraform (Plus tier). To be IaC compliant we are using tfe Terraform provider for all the HCP Terraform related resources. In order to configure certain private agent pools to be used by a certain number of workspaces, the agent pools must be created with the TFE provider.

Our choice for running the actual agents was using the V2 hcp-terraform-operator with the agentpool custom resource on AWS EKS, but unfortunately currently you can only create/manage new agent pools, not existing ones.

Our request would be to add logic/functionality to the agentpool custom resource to be able to use existing/blank agent pools.

Currently we are using a workaround with kubectl patch (but this is not a solid solution):

kubectl patch -n hcp-tf-ns agentpool agents-eks --type=merge --subresource status --patch 'status: {agentPoolID: apool-5739FGHfhryfj355sd, observedGeneration: 1}'

Potential YAML Configuration

spec:
  agentDeployment:
    replicas: 1
    spec:
      containers:
        - image: hashicorp/tfc-agent:1.20
          name: tfc-agent
  agentTokens:
    - name: agents-eks-token
  name: agents-eks
  organization: MytfOrg

  existingAgentPool: true    <--- in this case it should check based on the name of the agent pool if it exists or not and use the ID of that agent pool.

OR

  existingAgentPoolID: apool-5739FGHfhryfj355sd   <--- in this case the agent pool name is no longer required and just use the provided agent pool ID.

References

Community Note

• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.
• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request.
• If you are interested in working on this issue or have submitted a pull request, please leave a comment.

[dakky]

We have a similar issue:

For redundancy reasons we have 2 k8s Clusters, which should host agents for 1 pool.
When adding the second cluster we get the errors about the already existing pool.

Concerning the proposed approach of @zoli-opslogic:

imho its important to have a solution which does not require to hardcode the pool ID into the agentpool ressource as this will break any automation (fluxcd in our case) as the poolname is automtically generated and not known before