Notification Policies in YAML

The notificationPolicies section of devhelm.yml controls how incidents are routed to alert channels. Each policy has match rules that filter incidents and an escalation chain that defines notification steps.

Policy fields

Field	Type	Required	Description
`name`	string	Yes	Unique policy name
`enabled`	boolean	—	Whether the policy is active (default: `true`)
`priority`	integer	—	Evaluation priority (higher = evaluated first, default: `0`)
`matchRules`	object[]	—	Rules that filter which incidents trigger this policy
`escalation`	object	Yes	Escalation chain definition

Match rules

Match rules filter which incidents trigger the policy. If no rules are defined, the policy matches all incidents.

notificationPolicies:
  - name: Production Alerts
    matchRules:
      - type: monitor_tag_in
        values: [production]
      - type: region_in
        regions: [us-east, eu-west]

Field	Type	Required	Description
`type`	string	Yes	Rule type: `severity_gte`, `monitor_id_in`, `region_in`, `incident_status`, `monitor_type_in`, `service_id_in`, `resource_group_id_in`, `component_name_in`, `monitor_tag_in`
`value`	string	—	Single match value (e.g. severity for `severity_gte`)
`values`	string[]	—	Multiple match values (e.g. tag names for `monitor_tag_in`)
`monitorNames`	string[]	—	Monitor names for `monitor_id_in` (resolved to IDs at deploy time)
`regions`	string[]	—	Regions for `region_in`

Escalation chain

The escalation chain defines an ordered sequence of notification steps:

notificationPolicies:
  - name: Critical Alerts
    priority: 10
    escalation:
      steps:
        - channels: [Slack Alerts]
          delayMinutes: 0
        - channels: [PagerDuty On-Call]
          delayMinutes: 5
          requireAck: true
          repeatIntervalSeconds: 300
      onResolve: Slack Alerts
      onReopen: PagerDuty On-Call

Steps

Field	Type	Required	Description
`channels`	string[]	Yes	Alert channel names to notify
`delayMinutes`	integer	—	Minutes to wait before this step (default: `0`)
`requireAck`	boolean	—	Require acknowledgement before escalating
`repeatIntervalSeconds`	integer	—	Seconds between repeated notifications

Lifecycle hooks

Field	Type	Description
`onResolve`	string	Channel name to notify when incident resolves
`onReopen`	string	Channel name to notify when incident reopens

Priority and evaluation

Policies are evaluated in descending priority order. The first matching policy handles the incident. Use priority to ensure specific policies (e.g., production-only) take precedence:

notificationPolicies:
  - name: Production Critical
    priority: 20
    matchRules:
      - type: monitor_tag_in
        values: [production]
    escalation:
      steps:
        - channels: [PagerDuty On-Call]

  - name: Default Alerts
    priority: 0
    escalation:
      steps:
        - channels: [Slack Alerts]

Complete example

notificationPolicies:
  - name: Tiered Escalation
    enabled: true
    priority: 10
    matchRules:
      - type: monitor_tag_in
        values: [production, critical]
    escalation:
      steps:
        - channels: [Slack Alerts]
          delayMinutes: 0
        - channels: [PagerDuty On-Call]
          delayMinutes: 5
          requireAck: true
        - channels: [Ops Email]
          delayMinutes: 15
      onResolve: Slack Alerts

Overview

YAML Format

Deploy Workflow

Terraform Provider

CI/CD Patterns

Notification Policies in YAML

Policy fields

Match rules

Escalation chain

Steps

Lifecycle hooks

Priority and evaluation

Complete example

Next steps

Alert channels

Notification policies reference

​Policy fields

​Match rules

​Escalation chain

​Steps

​Lifecycle hooks

​Priority and evaluation

​Complete example

​Next steps

Alert channels

Notification policies reference

Policy fields

Match rules

Escalation chain

Steps

Lifecycle hooks

Priority and evaluation

Complete example

Next steps