Monitoring

Datadog Integration

Vellum supports real-time monitoring of platform events through Datadog integration. This allows organizations to leverage Datadog’s powerful monitoring, alerting, and business intelligence capabilities with their Vellum data.

Configuration

To set up Datadog integration:

  1. Navigate to your organization settings page in Vellum
  2. Locate the Datadog integration section
  3. Configure the integration with your desired event types

Supported Events

The following events are available for monitoring in Datadog:

CategoryEventDescription
Node Execution Eventsnode.execution.initiatedTriggered when a node begins execution
node.execution.fulfilledTriggered when a node successfully completes execution
node.execution.rejectedTriggered when a node execution fails
Workflow Execution Eventsworkflow.execution.initiatedTriggered when a workflow starts execution
workflow.execution.fulfilledTriggered when a workflow successfully completes
workflow.execution.rejectedTriggered when a workflow execution fails
Usage Calculation Eventsworkflow_execution.usage_calculation.fulfilledTriggered when workflow usage is successfully calculated
workflow_execution.usage_calculation.rejectedTriggered when workflow usage calculation fails
prompt_execution.usage_calculation.fulfilledTriggered when prompt usage is successfully calculated
prompt_execution.usage_calculation.rejectedTriggered when prompt usage calculation fails

Use Cases

The Datadog integration enables several key monitoring and analytics capabilities:

  1. Real-time Monitoring: Track workflow and node executions as they happen
  2. Performance Analytics: Analyze execution times and success rates
  3. Error Detection: Get immediate alerts when executions fail
  4. Usage Tracking: Monitor resource consumption and costs
  5. Business Intelligence: Create custom dashboards and reports using your Vellum data

Best Practices

When setting up your Datadog integration:

  1. Start with critical events first (e.g., execution failures)
  2. Set up appropriate alerting thresholds
  3. Create dashboards for commonly monitored metrics
  4. Use tags effectively to segment and analyze your data