Terraform lifecycle of a job that has ttl_seconds_after_finished set

[stefan-fast]

This is basically a copy of issue #2059, which unfortunately went stale due to a lack of activity. The issue is still relevant and affecting me.

Terraform version, Kubernetes provider version and Kubernetes version

Terraform version: v1.5.7
Kubernetes Provider version: v2.23.0
Kubernetes version: v1.28.9

Terraform configuration

terraform {
  required_providers {
    kubernetes = {
      source  = "hashicorp/kubernetes"
      version = ">= 2.18.0"
    }
  }

  required_version = ">= 1.4.0"
}

provider "kubernetes" {
  config_path = "~/.kube/config"
}


variable "test" {
  type    = number
  default = 1
}

resource "terraform_data" "kubernetes_job_control" {
  input = var.test
}

resource "kubernetes_job" "mock_job" {
  lifecycle {
    replace_triggered_by = [
      terraform_data.kubernetes_job_control
    ]
  }
  metadata {
    name = "mock-job"
  }
  spec {
    ttl_seconds_after_finished = 0
    template {
      metadata {}
      spec {
        container {
          name  = "mock-job"
          image = "hello-world"
        }
        restart_policy = "Never"
      }
    }
    backoff_limit = 0
  }
  wait_for_completion = true
}

Question

Hello,

I have a kubernetes_job that sets up some initial configuration in my db when the database is created, for that I did lifecycle { replace_triggered_by = [ my_database_resource.id ] }. The thing is, I don't want this job to stick around in the cluster so I've set ttl_seconds_after_finished=0. I thought that by setting the lifecycle block the job would only be applied whenever I need to recreate the database, but because the job deletes itself from cluster it is recreated in every terraform apply that I do. It looks like a bug to me but I think it would be better to ask for opinions here first.

The code I pasted here is a simplified terraform project to reproduce this problem in case someone wants to try it locally. The principle is the same, I have a variable controlling the lifecycle and the job's ttl_seconds_after_finished set to 0.