fix: prevent perpetual plan diff on serverless cluster computed attrributes

[bohnen]

What problem does this PR solve?

After creating Starter cluster, terraform plan always produces in-place update. Actually it changes noting but terraform plan returns 2 and cause confusion.

Terraform will perform the following actions:

  # tidbcloud_serverless_cluster.starter will be updated in-place
  ~ resource "tidbcloud_serverless_cluster" "starter" {
      ~ annotations             = {
          - "tidb.cloud/available-features" = "DISABLE_PUBLIC_LB,DELEGATE_USER"
          - "tidb.cloud/has-set-password"   = "false"
        } -> (known after apply)
      ~ labels                  = {
          - "tidb.cloud/organization" = "<org id>"
          - "tidb.cloud/project"      = "<project id>"
        } -> (known after apply)
      ~ state                   = "ACTIVE" -> (known after apply)
      ~ update_time             = "2026-07-03T05:22:36Z" -> (known after apply)
      ~ version                 = "v8.5.3" -> (known after apply)
        # (11 unchanged attributes hidden)
    }

Plan: 0 to add, 1 to change, 0 to destroy.

What is changed and how it works?

Add UseStateForUnknown plan modifiers to the computed-only attributes version, update_time, state, labels, and annotations. Without them these attributes were planned as "(known after apply)" on every no-op plan, producing a spurious in-place update. Matches the existing treatment of other computed attributes (create_time, cluster_id, etc.).

Ref: https://developer.hashicorp.com/terraform/plugin/framework/resources/plan-modification#common-use-case-attribute-plan-modifiers