Create custom monitors in Vercel Observability
Vercel Observability custom monitors can alert on individual metrics or metric ratios, using thresholds or anomaly detection. Builders can configure them in the UI or CLI and route alerts to several channels.
Custom monitors in Vercel Observability track a single metric or a ratio, with **anomaly detection or fixed thresholds**. They can cover AI Gateway costs, performance, failure rates, and Sandbox CPU time.
Builders can configure detection windows, filters, and grouping, then create monitors in the UI or with the **Vercel CLI**. Alerts can reach email, Slack, Vercel Inbox, push notifications, or team webhooks; Vercel Agent can receive investigation context.
Custom monitors in Vercel Observability track a single metric or a ratio, with **anomaly detection or fixed thresholds**. They can cover AI Gateway costs, performance, failure rates, and Sandbox CPU time. Builders can configure detection windows, filters, and grouping, then create monitors in the UI or with the **Vercel CLI**. Alerts can reach email, Slack, Vercel Inbox, push notifications, or team webhooks; Vercel Agent can receive investigation context. The feature is for Pro and Enterprise plans with Observability Plus. Scans cost **$0.0035 per GB**; built-in alerts remain included with Observability Plus.
Custom monitors add configurable anomaly or threshold alerts for operational metrics, including AI Gateway costs, failures, performance, and Sandbox CPU time. Vercel Agent can receive investigation context, linking detection to agent-assisted follow-up; availability and scan charges define practical adoption constraints.