Under resource - `google_monitoring_alert_policy` Add Severity option

[vxr222]

google_monitoring_alert_policy Under this resource i wanted to add option for severity
Ref. Link https://cloud.google.com/monitoring/api/ref_v3/rest/v3/projects.alertPolicies#severity
Terraform Link = https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/monitoring_alert_policy#severity

I am getting error

on main.tf line 60, in resource "google_monitoring_alert_policy" "dataflow_job_failed":
│   60:     severity = "ERROR"
│
│ An argument named "severity" is not expected here.

Module is below

resource "google_monitoring_alert_policy" "dataflow_job_failed" {
  project               = var.project_id
  severity = "ERROR"
  display_name          = "[Dataflow] Job Failed"
  notification_channels = var.notification_channels 
  combiner              = "OR"
  # user_labels           = var.user_labels
  conditions {
    display_name = "[Dataflow] Job Failed"
    condition_threshold {
      filter          = "metric.type=\"dataflow.googleapis.com/job/is_failed\" resource.type=\"dataflow_job\""
      duration        = var.duration
      comparison      = var.comparison
      threshold_value = var.threshold_value

      aggregations {
        per_series_aligner = "ALIGN_MAX"
        alignment_period   = "300s"
      }
      trigger {
        count = 1
      }
    }
  }
  alert_strategy {
    auto_close = var.auto_close_duration
  }
}

• Terraform community resources
• HashiCorp support (Terraform Enterprise customers)
• Google Cloud Platform Slack

[P1K3N]

Currently the severity available in the provider documentation it attaches the severity to the incident.
In the policy I can still see "No severity", so something is not working.
https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/monitoring_alert_policy#severity

[Lavieira]

should you label it as "bug"?

[cw-optima-ovo]

Just adding that I'm seeing this behaviour too, cannot add severity to a google_monitoring_alert_policy resource as per documentation:
https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/monitoring_alert_policy#severity

It is possible to add severity via the google cloud console and if exporting the JSON from the alert policy you can see there is a severity key available.
This PR should have solved the problem but even with the latest terraform version (1.7.5) it is not possible:
#16775

I would expect this to deploy but the severity flag is not recognised by terraform.

resource "google_monitoring_alert_policy" "test-severity" {
  display_name = "test-severity"
  severity = "WARNING"
  enabled      = true
  project      = "my-project"
  combiner     = "OR"
  conditions {
    display_name = "Error detected in logs"
    condition_matched_log {
      filter = "(resource.type = \"cloud_run_revision\"\nresource.labels.service_name = \"some-function\"\nresource.labels.location = \"europe-west1\")\n severity>=DEFAULT\""
    }
  }
  alert_strategy {
    notification_rate_limit {
      period = "600s"
    }
  }
}

This code produces the error

│ Error: Unsupported argument
│ 
│   on testfile.tf line 74, in resource "google_monitoring_alert_policy" "test-severity":
│   74:   severity = "WARNING"
│ 
│ An argument named "severity" is not expected here.

@vxr222, I would agree with @Lavieira this should be labelled as a bug

[luismendezescobar]

I'm having the same issue:
`resource "google_monitoring_alert_policy" "app_engine_read_queries" {
display_name = "Quota usage app engine read queries per minute on: ${var.app_name}"
enabled = var.alert_enabled_app_engine
combiner = "OR"
notification_channels = var.notification_channel

severity = "CRITICAL"

documentation {
content = "To resolve this issue, click [here](https://console.cloud.google.com/iam-admin/quotas?"
}
user_labels = {
configuration = "terraform"
}
conditions {
display_name = "Quota usage app engine read queries per minute on: ${var.app_name}"
condition_monitoring_query_language {
query = <<EOT
fetch consumer_quota|filter resource.project_id==${var.project_id}
EOT
duration = "60s"
trigger {
count = 1

  }
  
}

}
alert_strategy {
auto_close = "3600s"

}

}`

getting the error:
Error: Unsupported argument on ../../modules/alerts-quotes/main.tf line 7, in resource "google_monitoring_alert_policy" "app_engine_read_queries": 7: severity = "CRITICAL"

[raweber42]

same here

[AntonioEFX]

Me too...

[rileykarson]

This was added in 5.10.0: https://github.com/hashicorp/terraform-provider-google/releases/tag/v5.10.0. Could someone post terraform -v from the config directory, or capture provider debug logs when the error happens? Thanks!

I'm just clicking buttons while passing by and not digging deeper than this atm- I haven't run a config to confirm, just perused the code and nothing looked off- here's our configuration in acceptance tests, which confirms this should be working:

terraform-provider-google/google/services/monitoring/resource_monitoring_alert_policy_test.go

Lines 240 to 266 in 3af7bc9

	func testAccMonitoringAlertPolicy_basicCfg(alertName, conditionName, aligner, filter, severity string) string {
	return fmt.Sprintf(`
	resource "google_monitoring_alert_policy" "basic" {
	display_name = "%s"
	enabled = true
	combiner = "OR"
	conditions {
	display_name = "%s"
	condition_threshold {
	aggregations {
	alignment_period = "60s"
	per_series_aligner = "%s"
	}
	duration = "60s"
	comparison = "COMPARISON_GT"
	filter = "%s"
	threshold_value = "0.5"
	}
	}
	severity = "%s"
	}
	`, alertName, conditionName, aligner, filter, severity)
	}

[cw-optima-ovo]

@rileykarson
output from terraform -v

Terraform v1.8.2
on linux_amd64
+ provider registry.terraform.io/hashicorp/google v5.25.0

After now upgrading to the above versions, I can verify the severity flag is now working, example config here:

resource "google_monitoring_alert_policy" "test-severity" {
  display_name = "test-severity"
  severity = ""
  enabled      = true
  project      = "my-project"
  combiner     = "OR"
  conditions {
    display_name = "Error detected in logs"
    condition_matched_log {
      filter = "(resource.type = \"cloud_run_revision\"\nresource.labels.service_name = \"cloud_run_resource\"\nresource.labels.location = \"europe-west2\")\n severity>=ERROR"
    }
  }
  alert_strategy {
    notification_rate_limit {
      period = "600s"
    }
  }
}

[rileykarson]

Thanks! Just to confirm explicitly, it didn't work before upgrading but did work afterwards?

[cw-optima-ovo]

@rileykarson At the point of my comment from 15/03, this wasn't working but I didn't note the version of the provider (unhelpfully). I tested this in a fresh folder though so whatever the latest provider version would be I would likely have used though can't be certain.

In one of the configs I have running at the moment using provider v4.8.4, the severity flag doesn't work but it also doesn't exist in that provider version so that's to be expected.