ilert_schedule and ilert_team produce recurring changes

[brainbug95]

I set up an ilert_schedule like this

# recurring schedule
resource "ilert_schedule" "de_office-hours" {
  name     = "de-office-hours"
  timezone = "Europe/Berlin"
  type     = "RECURRING"
  schedule_layer {
    name      = "layer1"
    starts_on = "2025-01-01T00:00"
    user {
      id = ilert_user.hans_wurst.id
    }

    user {
      id = data.ilert_user.hein_bloed.id
    }

    rotation         = "P1D"
    restriction_type = "TIMES_OF_WEEK"

    restriction {
      from {
        day_of_week = "MONDAY"
        time        = "07:00"
      }
      to {
        day_of_week = "MONDAY"
        time        = "18:00"
      }
    }

    restriction {
      from {
        day_of_week = "TUESDAY"
        time        = "07:00"
      }
      to {
        day_of_week = "TUESDAY"
        time        = "18:00"
      }
    }

    restriction {
      from {
        day_of_week = "WEDNESDAY"
        time        = "07:00"
      }
      to {
        day_of_week = "WEDNESDAY"
        time        = "18:00"
      }
    }

    restriction {
      from {
        day_of_week = "THURSDAY"
        time        = "07:00"
      }
      to {
        day_of_week = "THURSDAY"
        time        = "18:00"
      }
    }

    restriction {
      from {
        day_of_week = "FRIDAY"
        time        = "07:00"
      }
      to {
        day_of_week = "FRIDAY"
        time        = "18:00"
      }
    }

  }
}

so far so good. But in any subsequent tofu apply this produces an alleged change:

OpenTofu will perform the following actions:

  # ilert_schedule.de_office-hours will be updated in-place
  ~ resource "ilert_schedule" "de_office-hours" {
        id        = "48153"
        name      = "de-office-hours"
        # (3 unchanged attributes hidden)

      ~ schedule_layer {
            name             = "layer1"
            # (3 unchanged attributes hidden)

          ~ restriction {
              ~ from {
                  ~ day_of_week = "THURSDAY" -> "MONDAY"
                    # (1 unchanged attribute hidden)
                }
              ~ to {
                  ~ day_of_week = "THURSDAY" -> "MONDAY"
                    # (1 unchanged attribute hidden)
                }
            }
          ~ restriction {
              ~ from {
                  ~ day_of_week = "WEDNESDAY" -> "TUESDAY"
                    # (1 unchanged attribute hidden)
                }
              ~ to {
                  ~ day_of_week = "WEDNESDAY" -> "TUESDAY"
                    # (1 unchanged attribute hidden)
                }
            }
          ~ restriction {
              ~ from {
                  ~ day_of_week = "FRIDAY" -> "WEDNESDAY"
                    # (1 unchanged attribute hidden)
                }
              ~ to {
                  ~ day_of_week = "FRIDAY" -> "WEDNESDAY"
                    # (1 unchanged attribute hidden)
                }
            }
          ~ restriction {
              ~ from {
                  ~ day_of_week = "MONDAY" -> "THURSDAY"
                    # (1 unchanged attribute hidden)
                }
              ~ to {
                  ~ day_of_week = "MONDAY" -> "THURSDAY"
                    # (1 unchanged attribute hidden)
                }
            }
          ~ restriction {
              ~ from {
                  ~ day_of_week = "TUESDAY" -> "FRIDAY"
                    # (1 unchanged attribute hidden)
                }
              ~ to {
                  ~ day_of_week = "TUESDAY" -> "FRIDAY"
                    # (1 unchanged attribute hidden)
                }
            }

            # (2 unchanged blocks hidden)
        }
    }

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

I had a similar issue with the ilert_team resource but here I managed to "fix" it by sorting my users within terraform code alphabetically. It would seem there is some sorting in place which is not accounted for within the terraform provider.

BR,
Christian

[STLVRTX]

Hello Christian, thanks for submitting this issue. I could reproduce your problem, we are already investigating the issue. I will post an update soon when we've decided on a solution.

[brainbug95]

Hi @STLVRTX

gleiches Problem bei "ilert_alert_action". Hier scheint die "alert_source" ebenfalls einer Sortierung zu unterliegen.

VG,
Christian

[brainbug95]

@STLVRTX any updates on this issue? It's messing up my drift detection pipeline.

[STLVRTX]

This issue needs some changes in the API, a fix in the provider itself is not possible as i cannot pin the order on apply. I've notified the engineering team again to tackle this soon.
I'm currently working on a new release for another feature, as soon as the API is changed I'll schedule another release.