tool_adapters/alertmgr.yaml

name: alertmgr
description: Alert and incident management system for monitoring, escalation, and on-call coordination
binary: alertmgr
auth:
  type: env_var
  key: ALERTMGR_API_KEY
commands:
  - name: alert list
    description: List alerts with optional filtering by severity and status
    args:
      - name: severity
        type: enum
        required: false
        description: Filter by alert severity
        values: ["critical", "warning", "info"]
      - name: status
        type: enum
        required: false
        description: Filter by alert status
        values: ["active", "acknowledged", "resolved"]
      - name: service
        type: string
        required: false
        description: Filter by originating service
      - name: limit
        type: int
        required: false
        description: Maximum number of alerts to return
    output_format: json
    side_effects: false
    example: "alertmgr alert list --severity critical --status active"

  - name: alert acknowledge
    description: Acknowledge an active alert
    args:
      - name: id
        type: string
        required: true
        description: Alert ID
      - name: comment
        type: string
        required: true
        description: Acknowledgment comment
    output_format: json
    side_effects: true
    example: "alertmgr alert acknowledge --id alt-001 --comment 'Investigating now'"

  - name: alert resolve
    description: Resolve an acknowledged or active alert
    args:
      - name: id
        type: string
        required: true
        description: Alert ID
      - name: resolution
        type: string
        required: true
        description: Resolution description
    output_format: json
    side_effects: true
    example: "alertmgr alert resolve --id alt-001 --resolution 'Scaled up instances, load normalized'"

  - name: rule create
    description: Create an alerting rule based on metric conditions
    args:
      - name: name
        type: string
        required: true
        description: Rule name (unique identifier)
      - name: condition
        type: string
        required: true
        description: "Alert condition expression (e.g. 'cpu_usage > 90 for 5m')"
      - name: severity
        type: enum
        required: true
        description: Alert severity when rule triggers
        values: ["critical", "warning", "info"]
      - name: notify
        type: string
        required: true
        description: Notification channel (e.g. slack channel, email, pagerduty)
      - name: description
        type: string
        required: false
        description: Human-readable rule description
    output_format: json
    side_effects: true
    example: "alertmgr rule create --name high-cpu --condition 'cpu_usage > 90 for 5m' --severity critical --notify '#ops-alerts'"

  - name: escalation create
    description: Create an escalation policy with ordered notification levels
    args:
      - name: name
        type: string
        required: true
        description: Escalation policy name
      - name: levels
        type: json
        required: true
        description: "Ordered escalation levels as JSON array [{\"delay_minutes\": int, \"notify\": str}]"
    output_format: json
    side_effects: true
    example: "alertmgr escalation create --name backend-escalation --levels '[{\"delay_minutes\": 0, \"notify\": \"on-call\"}, {\"delay_minutes\": 15, \"notify\": \"team-lead\"}, {\"delay_minutes\": 30, \"notify\": \"vp-eng\"}]'"

  - name: oncall show
    description: Show the current on-call schedule and active responder
    args:
      - name: team
        type: string
        required: false
        description: Team name (defaults to all teams)
    output_format: json
    side_effects: false
    example: "alertmgr oncall show --team backend"

  - name: incident create
    description: Create a new incident from one or more related alerts
    args:
      - name: title
        type: string
        required: true
        description: Incident title
      - name: severity
        type: enum
        required: true
        description: Incident severity
        values: ["critical", "warning", "info"]
      - name: alerts
        type: string
        required: false
        description: Comma-separated alert IDs to associate with this incident
      - name: commander
        type: string
        required: false
        description: Incident commander username
    output_format: json
    side_effects: true
    example: "alertmgr incident create --title 'Database outage' --severity critical --alerts alt-001,alt-002 --commander alice"