This product is not supported for your selected
Datadog site. (
).
このページは日本語には対応しておりません。随時翻訳に取り組んでいます。
翻訳に関してご質問やご意見ございましたら、
お気軽にご連絡ください。
Overview
Automations enable you to customize and extend incident management to fit your organization’s specific processes. Automatically trigger actions based on incident events such as severity changes, or state transitions.
Automations are powered by Datadog Workflow Automation and are included in your Incident Management billing at no additional cost.
Prerequisites
To create and manage automations, you must have the following permissions:
Workflows Write permissionIncident Settings Write OR Incident Notification Settings Write permission
To run automations on private incidents, use a user or service account with the Private Incidents Global Access permission. Without this permission, the automation cannot access incident data.
For more information on permissions, see Datadog Role Permissions.
Accessing automations
Automations are configured per incident type. To manage automations:
- Navigate to Incidents > Settings.
- Select an incident type from the list.
- Click the Automations tab.
From this page, you can view, create, enable, disable, and manage your automations.
Any user with Incident Settings Write or Incident Notification Settings Write permissions can toggle automations on or off. This is true even if they don't have edit access to the automation itself. Administrators can quickly disable problematic automations if needed.
Creating an automation
You can build automations entirely from the Incident Management settings UI. For more advanced workflows, open the automation in the Workflow Automation editor to access additional actions and logic capabilities.
When you click New Automation, you have two options for building your workflow:
Start with a blueprint
Blueprints provide pre-configured automation templates for common use cases, such as sending a Slack message to the incident channel. Using a blueprint is the fastest way to get started.
Choose an action
For custom processes, you can build an automation from scratch by starting with an individual action. You can choose from incident-specific actions or explore the full Datadog Action Catalog, which contains thousands of integrations.
Configuring triggers and conditions
Trigger types
Select when your automation should run:
| Trigger Type | Description |
|---|
| When the incident is declared | Runs once when an incident is first declared and meets the defined conditions. |
| When the incident is declared or updated | Runs when an incident is declared or when any field changes that cause the incident to meet the conditions. |
| On a schedule | Runs repeatedly on a per-incident basis (for example, every 10 minutes for each active incident that meets the conditions). Useful for periodic reminders and status checks. |
Conditions
Define conditions to specify which incidents trigger the automation. Conditions are based on incident attributes such as Severity, State, Teams, or other custom property fields.
- Logic within a row: Selecting multiple values for a single property (like
SEV-1 and SEV-2) uses OR logic. - Logic across rows: Adding multiple property filters uses
AND logic.
Example: Set conditions for severity:SEV-1, severity:SEV-2, and summary:is empty. The automation runs when the incident is (SEV-1 OR SEV-2) AND the summary is empty.
Building automation workflows
Automations use the Datadog Workflow Automation engine. Each automation is a workflow that can include multiple actions and logic steps.
Using incident data
Automations have access to all incident data through the incident context variable, which includes:
incident.id: The incident’s unique identifierincident.attributes: All incident attributes (severity, state, title, custom fields, and more)incident.fieldDiffs: A list of fields that changed (for update triggers)
Use these variables in your automation actions by referencing them with curly braces, such as {{ incident.id }}.
Configuring actions
Each action in your automation requires configuration. For example, to send a message to an incident’s Slack channel:
- Add the Get incident Slack channel action.
- Set the input parameter to
{{ incident.id }}. - Add the Send Slack message action.
- Configure the message content using incident variables.
The workflow editor provides autocomplete for available variables and validates your configuration.
Testing automations
There are two ways to test your automations:
Option 1: Declare a test incident
- Enable test incidents in your incident settings.
- Declare a test incident that matches your automation’s conditions.
- View the automation execution in the incident timeline.
Option 2: Test from an existing incident
- Open the automation in the workflow editor.
- Click the Run button.
- Select Test from incident.
- Choose an existing incident to simulate the trigger.
This populates the incident context variable with data from the selected incident without actually triggering the automation for that incident.
Viewing automation executions
From the incident timeline
Every automation execution appears in the incident timeline. Timeline entries include:
- The automation name
- Execution timestamp
- Link to the detailed execution view
- Execution status (success or failure)
You can filter the timeline to show only automation executions or exclude them entirely.
From execution history
To view all executions of an automation:
- Open the automation.
- Click Execution in the workflow editor.
The execution history shows:
- All input parameters and their values
- The
incident context data - The
fieldDiffs showing what changed - Step-by-step execution results
- Any errors or failures
Permissions and access control
Edit access
By default, only the automation creator can edit an automation. To grant edit access to others:
- Open the automation.
- Click Edit Access.
- Add users or service accounts.
Granting edit access allows others to use the Datadog API as you or as the service account. Use service accounts for shared automations to avoid issues when users leave the organization.
Service accounts
Using a service account to run automations provides several benefits:
- Automations continue running if the creator leaves the organization
- Better separation of duties and access control
- Clearer audit trails
To use a service account:
- Open the automation.
- Click Run as Service Account.
- Create a new service account with appropriate roles or select an existing one.
You must have the Service Account Write permission to configure service accounts for automations.
Private incidents
Automations can run on private incidents with the following considerations:
Required permissions
To run automations on private incidents, use a user or service account with the Private Incidents Global Access permission. Without this permission, the automation cannot access incident data.
Security considerations
By default, execution history (including private incident data) is visible to anyone in your organization. To run automations on private incidents securely:
- Use a service account with
Private Incidents Global Access permission. - Restrict viewer access to only users who should see private incident data.
Differences from notification rules
Both automations and notification rules can respond to incident events, but they serve different purposes:
| Feature | Automations | Notification Rules |
|---|
| Purpose | Execute complex workflows and integrations | Send notifications to stakeholders |
| Triggers | Declared, updated, or scheduled | Declared or updated |
| Actions | Access to full Datadog Action Catalog | Limited to notification channels |
| Complexity | Multi-step workflows with logic | Single notification per rule |
| Cost | Included in Incident Management | Included in Incident Management |
Use notification rules for straightforward notifications and automations for complex, multi-step processes.
Use cases and examples
Use the following examples to help you build your own incident automations.
Trigger: When declared or updated
Condition: Severity is SEV-1 or SEV-2
Actions:
- Detect when teams field changes
- Add new teams to the incident teams list
- For all users in the team, invite them to the incident slack channel
Access the blueprint in Datadog.
Trigger: On a schedule (every 30 minutes)
Condition: Severity is SEV-1 or SEV-2, State is Active or Stable
Actions:
- Check time since last update
- Send Slack reminder if > 30 minutes
- Prompt commander to update incident status
Access the blueprint in Datadog.
Further reading