Root cause

The monitor shows No Data when you graph the metric. No data appears because there is no data.

Verify whether data exists

1. Identify the metric (such as AWS metric or custom metric).
2. Graph it in a notebook or dashboard.
3. Scope the graph to the same timeframe as the monitor.
4. Check if data points appear.

Understanding (no groups found) vs. No Data

While they look similar, (no groups found) is specific to multi alert monitors (grouped by a tag). It indicates that the monitor has no active entities to evaluate, whereas No Data refers to the metric itself being empty and is controlled by your monitor’s If data is missing settings.

(no groups found) typically occurs in the following cases:

• New Setup: The monitor has received no data for any groups since creation.

• Aged Out: Previously active groups stopped reporting and have been cleared from the monitor’s memory. Different types of groups have different retention periods:

  • Standard Monitors: Groups age out after 24 hours.
  • Host/Service Checks: Groups age out after 48 hours.
  • Custom: This can be extended through the Missing data options setting (If a group stops reporting data, resolve after X hours).

Root cause

Data exists but doesn’t appear consistently or arrives late.

Verify data frequency

1. Graph the metric over 2-3 times the evaluation timeframe.
   • Example: For a 5-minute evaluation window, graph the last 15 minutes.
2. Look for gaps between data points or inconsistent reporting intervals, which indicate sparse or delayed metrics.

Possible solutions

• In Metric monitor Advanced Options, enable the “Do not require a full window of data for evaluation” setting.
• Increase the evaluation window to capture more data points.
• Add an evaluation delay to account for data ingestion delays.
• For aggregations like avg, min, and max, consider a larger timeframe.

To learn how to apply these changes, see Adjusting No Data alerts for metric monitors.

Root cause

The monitor shows No Data even though data exists. If your query contains a rollup function, this could be causing No Data.

Why this happens

Rollups aggregate metric data into buckets. If the monitor doesn’t have at least one complete bucket, it reports No Data. Additionally, rollup intervals are aligned to UNIX time, not to the start and end of monitor queries. A monitor may evaluate an incomplete rollup interval containing only a small sample of data, which can result in No Data.

For example, if a monitor has a 4-minute rollup and a 20-minute evaluation window, it produces one data point every 4 minutes, leading to a maximum of 5 data points within the window. If the Require Full Window option is enabled, the evaluation may result in “No Data” because the window is not fully populated.

Solution

If you use rollups in your monitor query, add an evaluation delay at least equal to the rollup period. This ensures the monitor has at least one bucket of data to evaluate. For example, a 1-hour rollup requires a 3600-second evaluation delay.

Root cause

The monitor shows No Data, but data exists in dashboards. The metric is missing one or more tags that the monitor groups by.

Why this happens

When a metric is missing one or more of the tags your monitor query groups by, Datadog assigns N/A to those tags. In dashboards, N/A values are displayed automatically, but monitors use AND logic — all specified tags must be present on the metric. If any required tag is N/A, the monitor reports No Data. For example, a query scoped by env:production won’t match metrics where the env tag is N/A.

Identify N/A tag values

Confirm that N/A tag values are causing No Data in your monitor:

1. Create a notebook or dashboard with your monitor query.
2. Look for groups with N/A tag values.

Solution

Ensure that all tags specified in your monitor query are present on your metric data. If any tag is missing (shows as N/A), update your metric submission to include all required tags, or adjust your monitor query to group or scope only by tags that are always available.

Root cause

Unlike standard monitors, Anomaly and Forecast monitors depend on historical data to evaluate. Without enough history, they cannot produce a result and report No Data.

Why this happens

• Anomaly monitors require at least 3 seasons (repeating cycles, such as daily or weekly) of historical data to establish a baseline.
• Forecast monitors need varying amounts of historical data depending on the length of the forecast period.

If your query contains a rollup, see Query with a rollup shows No Data.

Solution

Allow time for the metric to accumulate more data.

Root cause

A monitor reports No Data when any query in its arithmetic expression returns NaN.

Why this happens

If any part of an arithmetic expression is NaN (Not a Number, which can happen in cases like trying to divide by zero), the entire expression becomes NaN and won’t evaluate. For example:

• Expression: a + b.
• If a is NaN, then a + b is NaN.
• The monitor won’t evaluate.

Identify the NaN query

Graph each part of the arithmetic expression separately to confirm which query is returning NaN:

1. Create a notebook or dashboard.
2. Graph each part of the arithmetic expression separately.
3. Identify which query returns NaN.
4. If at least one query meets No Data criteria, the monitor behaved as expected.