What happens if removed node pools are NOT from the end of list?

[knuurr]

Hi!

I’m trying to make sure I fully understand the correct (and safe) behavior around removing nodes and nodepools in kube.tf config, and what the consequences are if the suggested practices aren’t followed.

From the documentation:

Once the cluster is up, you can change any nodepool count and even set it to 0 (except the first control-plane nodepool, which must have at least 1). You can rename a nodepool (if the count is 0), but should not remove a nodepool from the list once the cluster is up.
The reason is related to how subnets and IPs are allocated — only nodepools at the end of the list can be safely removed.
You can freely add nodepools at the end, and for each nodepool you can scale the node count up or down (draining nodes before removal is recommended to avoid cluster issues).

My questions for clarification

• How can I detect or monitor if this issue already affects my hetzner deployment?
  • Are there observable symptoms and how can I detect them, either in hetzner web console, or through cli, or other?
• What exactly happens if the recommended behavior (as in how node pools are being removed i.e from the middle of the list) is not followed?
• What’s the recommended way to take corrective action if this does occur?

I must say I did not see this piece of info initially in docs and I am sure there were certain node pools (either for CA or static ones) that I did remove, and which were not at the end of the list, but for example in the middle, but definitely not last as outlined.