Adjusting No Data alerts for metric Monitors

Overview

When a metric monitor tracks a metric that doesn’t always report at the same frequency, or arrives with a timestamp in the past (such as metrics from AWS integrations), you may receive No Data alerts even when data appears in Datadog. Adjust the following settings in your monitor’s Advanced Options to reduce false No Data alerts.

Add an evaluation delay

The evaluation delay pushes the monitor’s evaluation window back in time, giving delayed metrics time to arrive in Datadog before the monitor evaluates.

AWS monitor config

This setting is recommended for:

  • AWS and crawler-based metrics: Set to at least 900 seconds (15 minutes) to account for CloudWatch delays.
  • Other cloud provider metrics: See Cloud metric delays for provider-specific recommendations.

Disable Require a Full Window of Data

The Require a Full Window of Data setting controls whether the monitor waits for a complete evaluation window before alerting. For sparse or backfilled metrics, requiring a full window can cause the monitor to skip evaluations or report No Data.

Monitor Advanced Options - Require a Full Window of Data setting in the Datadog UI
  • Do not require (recommended for sparse or delayed metrics): The monitor evaluates on partial data, reducing false No Data alerts.
  • Require: The monitor waits for a full window of data before evaluating. This is appropriate for metrics reported by the Datadog Agent at a consistent frequency.

Install the Datadog Agent on cloud hosts

Cloud metrics have varying delays depending on the provider. To receive metrics with a smaller delay, install the Datadog Agent directly on your cloud hosts. See installing the Datadog Agent on cloud instances for more information.

Further reading