Send Alerts to Opsgenie

Configure Edge Delta to create Opsgenie alerts using the Alert API for on-call management and incident response.

5 minute read

Overview

This guide demonstrates how to create Opsgenie alerts from EdgeDelta using the Alert API. Opsgenie is an incident management platform that routes alerts to the right on-call responders and enables effective incident response.

Prerequisites

Opsgenie Account: Access to create API integrations
API Key: A GenieKey from an Opsgenie integration
EdgeDelta Pipeline: A configured threshold node or monitor generating signals

Pipeline Flow

flowchart LR classDef icon-metrics fill:#E7EEFC,stroke:#2563EB,color:#0F2169; classDef icon-alert fill:#FDEAD7,stroke:#EA580C,color:#7C2D12; classDef icon-destination fill:#FCEADB,stroke:#EA580C,color:#7C2D12; classDef icon-notify fill:#FCE7E7,stroke:#DC2626,color:#7C1D1D; Metrics[" Metric Source"] Threshold[" Threshold Node"] Webhook[" Webhook Destination"] Opsgenie[" Opsgenie"] Metrics --> Threshold Threshold --> Webhook Webhook --> Opsgenie class Metrics icon-metrics; class Threshold icon-alert; class Webhook icon-destination; class Opsgenie icon-notify;

Diagram preview unavailable

View diagram source

flowchart LR
classDef icon-metrics fill:#E7EEFC,stroke:#2563EB,color:#0F2169;
classDef icon-alert fill:#FDEAD7,stroke:#EA580C,color:#7C2D12;
classDef icon-destination fill:#FCEADB,stroke:#EA580C,color:#7C2D12;
classDef icon-notify fill:#FCE7E7,stroke:#DC2626,color:#7C1D1D;
    Metrics["<span class='ph ph-chart-line-up'></span> Metric Source"]
    Threshold["<span class='ph ph-funnel'></span> Threshold Node"]
    Webhook["<span class='ph ph-paper-plane-tilt'></span> Webhook Destination"]
    Opsgenie["<span class='ph ph-bell-ringing'></span> Opsgenie"]
    Metrics --> Threshold
    Threshold --> Webhook
    Webhook --> Opsgenie
class Metrics icon-metrics;
class Threshold icon-alert;
class Webhook icon-destination;
class Opsgenie icon-notify;

Configuration

Step 1: Create an Opsgenie API Integration

Log in to your Opsgenie account
Navigate to Settings → Integration List
Search for API and select it
Click Add
Configure the integration name and team assignment
Copy the API Key (GenieKey)
Save the integration

Step 2: Store the API Key

Store the API key as an environment variable:

export OPSGENIE_API_KEY="your-opsgenie-api-key"

Step 3: Configure the Webhook Node

Create a webhook node that sends alerts to Opsgenie:

- name: opsgenie_alerts
  type: webhook_output
  endpoint: https://api.opsgenie.com/v2/alerts
  headers:
  - header: Content-Type
    value: "application/json"
  - header: Authorization
    value: "GenieKey ${OPSGENIE_API_KEY}"
  suppression_window: 20m
  payload: |
    {
      "message": "{{ .item.signal.title }}",
      "alias": "{{ .item.signal.signal_id }}",
      "description": "{{ .item.signal.description }}\n\nHost: {{ index .item.resource \"host.name\" }}\nSource: {{ index .item.resource \"ed.source_name\" }}\nValue: {{ .item.signal.value }}\nThreshold: {{ .item.signal.threshold_value }}",
      "tags": [
        "EdgeDelta",
        "{{ index .item.resource \"ed.source_type\" }}",
        "{{ index .item.resource \"host.name\" }}"
      ],
      "details": {
        "signal_id": "{{ .item.signal.signal_id }}",
        "metric_name": "{{ .item.signal.name }}",
        "current_value": "{{ .item.signal.value }}",
        "threshold": "{{ .item.signal.threshold_value }}",
        "threshold_type": "{{ .item.signal.threshold_type }}",
        "host": "{{ index .item.resource \"host.name\" }}",
        "source_name": "{{ index .item.resource \"ed.source_name\" }}",
        "source_type": "{{ index .item.resource \"ed.source_type\" }}",
        "agent_tag": "{{ index .item.resource \"ed.tag\" }}",
        "timestamp": "{{ .item.timestamp }}"
      },
      "priority": "P2"
    }

Step 4: Connect to a Threshold Node

nodes:
  - name: memory_threshold
    type: threshold
    filter: item.name == "system.memory.usage"
    condition: value > 85

  - name: opsgenie_alerts
    type: webhook_output
    # ... configuration from above ...

links:
  - from: memory_threshold
    to: opsgenie_alerts

Alert API Payload Reference

Required Fields

Field	Description	Example
`message`	Alert title (max 130 chars)	`{{ .item.signal.title }}`

Recommended Fields

Field	Description	Usage
`alias`	Unique identifier for deduplication	Signal ID
`description`	Detailed information (max 15000 chars)	Alert details and context
`priority`	Alert priority: `P1` to `P5`	`P2` for standard alerts

Optional Fields

Field	Description	Usage
`responders`	Teams, users, or escalations to notify	Team assignment
`visibleTo`	Teams/users who can see the alert	Access control
`actions`	Custom actions for the alert	`["Restart Service", "Check Logs"]`
`tags`	Labels for categorization	`["EdgeDelta", "production"]`
`details`	Custom key-value properties	Metric details
`entity`	Domain/entity the alert relates to	Service name
`source`	Source of the alert	`EdgeDelta`
`note`	Additional note for the alert	Context for responders

Dynamic Priority Mapping

Map priority based on metric severity:

payload: |
  {
    "message": "{{ .item.signal.title }}",
    "alias": "{{ .item.signal.signal_id }}",
    "priority": "{{ if gt .item.signal.value (mul .item.signal.threshold_value 3) }}P1{{ else if gt .item.signal.value (mul .item.signal.threshold_value 2) }}P2{{ else if gt .item.signal.value .item.signal.threshold_value }}P3{{ else }}P4{{ end }}",
    ...
  }

Priority mapping:

P1: Critical - value exceeds 3x threshold
P2: High - value exceeds 2x threshold
P3: Moderate - value exceeds threshold
P4: Low - informational

Routing to Specific Teams

Route alerts to specific responders based on source:

payload: |
  {
    "message": "{{ .item.signal.title }}",
    "responders": [
      {
        "type": "team",
        "name": "{{ if contains (index .item.resource \"ed.source_name\") \"database\" }}Database Team{{ else }}Infrastructure Team{{ end }}"
      }
    ],
    ...
  }

Or use static team assignments:

payload: |
  {
    "responders": [
      {"type": "team", "name": "Infrastructure"},
      {"type": "user", "username": "oncall@company.com"}
    ],
    ...
  }

Adding Custom Actions

Include action buttons in the Opsgenie alert:

payload: |
  {
    "message": "{{ .item.signal.title }}",
    "actions": [
      "View in EdgeDelta",
      "Restart Service",
      "Acknowledge",
      "Add Note"
    ],
    ...
  }

Auto-Close Alerts

Create a separate webhook to close alerts when metrics return to normal:

- name: opsgenie_close
  type: webhook_output
  endpoint: https://api.opsgenie.com/v2/alerts/{{ .item.signal.signal_id }}/close?identifierType=alias
  headers:
  - header: Content-Type
    value: "application/json"
  - header: Authorization
    value: "GenieKey ${OPSGENIE_API_KEY}"
  payload: |
    {
      "note": "Alert auto-closed by EdgeDelta at {{ .item.timestamp }}. Metric returned to normal."
    }

Note: The alias must match the original alert’s alias for the close operation to work.

EU Region Configuration

If your Opsgenie account is in the EU region, use the EU endpoint:

endpoint: https://api.eu.opsgenie.com/v2/alerts

Testing

Test with Opsgenie’s API Console

Go to Settings → Integration List → Your API integration
Use the Test section to verify connectivity
Check the Logs tab for request/response details

Test with a Local Endpoint

Verify your payload structure before connecting to Opsgenie:

endpoint: http://localhost:8080/webhook-test

Expected Response

Successful alert creation returns:

{
  "result": "Request will be processed",
  "took": 0.302,
  "requestId": "43a29c5c-3dbf-4fa4-9c26-f4f71023e120"
}

Troubleshooting

Issue	Solution
`401 Unauthorized`	Verify GenieKey and `Authorization` header format
`403 Forbidden`	Check API integration permissions
`422 Unprocessable Entity`	Validate payload fields and character limits
Duplicate alerts	Use consistent `alias` values
Alerts not auto-closing	Verify alias matches and endpoint format

Character Limits

message: 130 characters max
alias: 512 characters max
description: 15000 characters max
tags: 20 tags max, 50 characters each

Best Practices

Use aliases for deduplication: Prevent duplicate alerts with consistent signal IDs
Set appropriate priorities: Reserve P1 for genuine critical issues
Include context in details: Custom details help responders investigate faster
Tag consistently: Use tags for filtering and reporting
Configure suppression windows: 20-30 minutes works well for most alerts
Route to the right teams: Use responders to ensure proper escalation
Test in non-production first: Validate payloads before production deployment