Create an alert

Create a new alert to monitor customer spending, balances, and billing metrics in real-time. Metronome’s alert system provides industry-leading speed with immediate evaluation capabilities, enabling you to proactively manage customer accounts and prevent revenue leakage.

This endpoint creates configurable alerts that continuously monitor various billing thresholds including spend limits, credit balances, commitment utilization, and invoice totals. Alerts can be configured globally for all customers or targeted to specific customer accounts. Custom fields can be used on certain alert types to further target alerts to groups of customers.

Use this endpoint to:

• Proactively monitor customer spending patterns to prevent unexpected overages or credit exhaustion
• Automate notifications when customers approach commitment limits or credit thresholds
• Enable real-time intervention for accounts at risk of churn or payment issues
• Scale billing operations by automating threshold-based workflows and notifications

Key response fields:

A successful response returns a CustomerAlert object containing:

• The alert configuration with its unique ID and current status
• The customer’s evaluation status (ok, in_alarm, or evaluating)
• Alert metadata including type, threshold values, and update timestamps

Usage guidelines:

• Immediate evaluation: Set evaluate_on_create : true (default) for instant evaluation against existing customers
• Uniqueness constraints: Each alert must have a unique uniqueness_key within your organization. Use release_uniqueness_key : true when archiving to reuse keys
• Alert type requirements: Different alert types require specific fields (e.g., billable_metric_id for usage alerts, credit_type_id for credit-based alerts)
• Webhook delivery: Alerts trigger webhook notifications for real-time integration with your systems. Configure webhook endpoints before creating alerts
• Performance at scale: Metronome’s event-driven architecture processes alert evaluations in real-time as usage events stream in, unlike competitors who rely on periodic polling or batch evaluation cycles