We have a public cloud setup entirely done as IaC with Terraform, the main element being a Kubernetes cluster.
We also have an ovh_cloud_project_loadbalancer with listeners configured. These listeners have their Ingress pool members also defined in Terraform. There's a member for each node in the cluster.
Now, when we try to scale a cluster with Terraform (i.e. add a node), each listener pool is supposed to get a new member. The OVH provider then wants to destroy the load balancer and recreate it which will disrupt our services. The only solution, it seems, would be to not manager the pool members with Terraform but instead configure them manually after each scale operation. This kind of defeats the purpose of IaC, though.
What are we missing? How are we supposed to handle this?
Happy for any insights.
Thanks,
Raphael