How To Terraform Datadog Monitors with Dynamic Thresholds

Marshall Davis

7/17/24, 3:55 PM

I'm attempting to create a number of resources from the DataDog provider. I'm hoping to have to define as little as possible for each resource. For many properties there is some sensible default. I am having difficulty in deciding how to handle monitor thresholds, specifically that some may be optional.

resource "datadog_monitor" "monitor" {
  for_each = {
    faulty_deploy = {
      message    = "A deployment failed."
      name       = "Deployment Failure"
      query      = "someLongQuery"
      type       = "alert"
      thresholds = {critical = "0"}
    }
  }

  message            = each.value["message"]
  query              = each.value["query"]
  name               = each.value["name"]
  type               = each.value["type"]
  escalation_message = coalesce(each.value["escalation_message"], "")
  evaluation_delay   = coalesce(each.value["evaluation_delay"], "0")
  include_tags       = coalesce(each.value["include_tags"],"true")
  
  dynamic "monitor_thresholds" {
    for_each = each.value["thresholds"]
    iterator = threshold
    content {
      # Dynamically adding these properties is the issue.
      "${threshold.key}" = threshold.value
    }
  }

  # These and more would use coalesce to set defaults.
  new_group_delay      = "60"
  notify_audit         = "false"
  on_missing_data      = "default"
  priority             = "0"
  renotify_interval    = "0"
  renotify_occurrences = "0"
  require_full_window  = "false"
  tags                 = []
  timeout_h            = "0"
}

Given that some thresholds may or may not be set how does one add dynamic properties in this way? Should they all always exist as 0 or "" and be optionally overwritten?

1 + 0

terraform

datadog

Score:0

Server

Marshall Davis

7/18/24, 3:29 PM

In short, you can't.

You can, however, include the monitor_thresholds nested schema with defaults in all cases. Given the same format as listed in the question:

monitor_thresholds {
  critical = lookup(each.value["thresholds"], "critical", "")
  warning = lookup(each.value["thresholds"], "warning", "")
}

The lookup function allows for a missing key and provides for a default if so. This can be expanded with a conditional for further optional sections.

monitor_threshold_windows {
  recovery_window = lookup(each.value, "threshold_windows", null) != null ? lookup(each.value["threshold_windows"], "recovery_window", "") : ""
}

+ 0

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: How To Terraform Datadog Monitors with Dynamic Thresholds

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.