Bug: Stackit Loadbalancer is recreated after acl change

[jojomo96]

Description

When i change an acl in terraform, the hole loadbalancer is recreated => 10 - 20 min downtime of this loadbalancer for a simple change like acl.

Steps to reproduce

Create a loadbalancer in tf, change the acl to something else ⇒ apply

 resource "stackit_loadbalancer" "slb" {
      ~ id                                = "..." -> (known after apply)
        name                              = "..."
      ~ options                           = { # forces replacement
          ~ acl                  = [ # forces replacement
              + "new acl/20",
              + "new acl/20",
                # (179 unchanged elements hidden)
            ]
          ~ observability        = { # forces replacement
              ~ logs    = {
                  + credentials_ref = (known after apply)
                  + push_url        = (known after apply)
                } -> (known after apply)
              ~ metrics = {
                  + credentials_ref = (known after apply)
                  + push_url        = (known after apply)
                } -> (known after apply)
            } -> (known after apply) # forces replacement
            # (1 unchanged attribute hidden)
        }
      ~ plan_id                           = "p10" # forces replacement -> (known after apply) # forces replacement
      + private_address                   = (known after apply)
      ~ security_group_id                 = "..." -> (known after apply)
        # (7 unchanged attributes hidden)
    }

module.stackit.stackit_loadbalancer.slb["..."]: Still creating... [19m0s elapsed]
module.stackit.stackit_loadbalancer.slb["..."]: Still creating... [19m10s elapsed]
module.stackit.stackit_loadbalancer.slb["..."]: Still creating... [19m20s elapsed]
module.stackit.stackit_loadbalancer.slb["..."]: Still creating... [19m30s elapsed]
module.stackit.stackit_loadbalancer.slb["..."]: Still creating... [19m40s elapsed]
module.stackit.stackit_loadbalancer.slb["..."]: Creation complete after 19m45s [id=...]

Actual behavior

The Loadbalancer is recreated on acl change

Expected behavior

The Loadbalancer just updated the acl, similar to the api

Environment

• OS:
• Terraform version (see terraform --version): OpenTofu v1.11.5
• Version of the STACKIT Terraform provider: v0.88.0

[marceljk]

Hi @jojomo96,

thanks for reporting this issue. The field cannot be updated in-place at the moment in our Terraform provider and will always trigger a replace. Probably it wasn't possible in the past to update it via the API, so we had to trigger a replace.
Please check always carefully the output of the Terraform provider to prevent unexpected replacements.

I create a ticket in our internal backlog to fix this soon.

Best regards
Marcel from the STACKIT Developer Tools Team