Terraform module to convert Prometheus Alerting rules to Grafana-managed alert rules
There are plenty of apps (mostly out of CNCF's ecosystem) where the vendor or the community provides monitoring dashboards and alerts. Dashboards are normally provided as a JSON file which can be loaded into Grafana. Alerts are mostly provided as Prometheus Alerting rules.
There are users who already operate a Grafana instance or use a managed Grafana instance from a cloud provider (Grafana Cloud, Amazon Managed Grafana, Azure Managed Grafana, etc.). Why not using this Grafana instance for the alerting?
The problem is that Grafana's unified alerting uses another format for the alert definition but the concept with labels, annotations (provide description and runbook URLs) is almost identical. This module allows you to reuse the Prometheus Alerting rules and configure them inside Grafana.
module "cert_manager_rules" {
source = "github.com/mkilchhofer/terraform-grafana-prometheus-alerts"
prometheus_alerts_file_path = file("/path/to/alerts/cert-manager.yaml")
folder_uid = grafana_folder.test.uid
datasource_uid = grafana_data_source.prometheus.uid
}
- Grafana 8.0+ (Unified alerting)
- Defining multiple alerts with the same name is not supported in Grafana
The alerts defined as Prometheus Alerting rules can be overridden without changing the input file.
If your Prometheus Alerting rules YAML contains an alert alert: TooManyWriteErrors
, you can override it like this:
module "cert_manager_rules" {
source = "github.com/mkilchhofer/terraform-grafana-prometheus-alerts"
prometheus_alerts_file_path = file("/path/to/alerts/cert-manager.yaml")
folder_uid = grafana_folder.test.uid
datasource_uid = grafana_data_source.prometheus.uid
# Overrides per alert
overrides = {
"TooManyWriteErrors" = {
alert_threshold = 1
is_paused = true
labels = {
mycustomlabel = "foobar" # Add an extra label
severity = "notify" # Override severity
}
}
}
}
Every alert supports the following overrides:
Override parameter | Type | Description |
---|---|---|
alert_threshold |
number |
Threshold of the Grafana alert. Defaults to 0 |
exec_err_state |
string |
Describes what state to enter when the rule's query is invalid and the rule cannot be executed. Options are OK , Error , KeepLast , and Alerting . Defaults to Error . |
expr |
string |
Use a custom PromQL expression for this Grafana alert. |
is_paused |
bool |
Sets whether the alert should be paused or not. Defaults to false . |
no_data_state |
string |
Describes what state to enter when the rule's query returns No Data. Options are OK , NoData , KeepLast , and Alerting . Defaults to OK . |
annotations |
map(string) |
Extra annotations to add. It is also possible to override already defined annotations like runbook_url . |
labels |
map(string) |
Extra labels to add. It is also possible to override already defined labels like severity . |
Name | Version |
---|---|
grafana | ~> 3.2 |
Name | Description | Type | Default | Required |
---|---|---|---|---|
datasource_uid | The UID of the Grafana datasource being queried with the expressions inside the Alerting rule file | string |
n/a | yes |
default_evaluation_interval_duration | How often is the rule evaluated by default. (When not defined inside your Alerting rules file) | string |
"5m" |
no |
disable_provenance | Allow modifying the rule group from other sources than Terraform or the Grafana API. | bool |
false |
no |
folder_uid | The UID of the Grafana folder that the alerts belongs to. | string |
n/a | yes |
org_id | The Organization ID of of the Grafana Alerting rule groups. (Only supported with basic auth, API keys are already org-scoped) | string |
null |
no |
overrides | Overrides per Alert rule | map(object({ |
{} |
no |
prometheus_alerts_file_path | Path to the Prometheus Alerting rules file | string |
n/a | yes |
No outputs.