Overview

Datadog and OpenTelemetry (OTel) use different naming conventions for integration metrics. This guide explains how to combine metrics from both systems in a single query using Datadog’s equiv_otel function.

Challenges when combining metrics

When working with both Datadog and OTel metrics, two main challenges arise. Let’s examine these using NGINX connection monitoring as an example:

Different naming conventions

Datadog and OTel handle the same measurements differently:

• Datadog: nginx.net.connections (a specific metric for active connections)
• OTel: nginx.connections_current (captures all connection states in a single metric)
  • Requires filtering with state:active to match Datadog’s active connections metric

Aggregation limitations

Simply combining separate metric queries can lead to incorrect results. For example, if you try to combine these queries:

avg:nginx.net.connections
avg:nginx.connections_current{state:active}

You get an average of averages, not the true average across all timeseries. This happens because traditional metrics functions combine the results of separate queries rather than treating the data as a single metric.

Combining metrics with the equiv_otel function

The equiv_otel function automatically combines equivalent Datadog and OTel metrics in a single query. It:

• Automatically handles metric name translation
• Properly aggregates all timeseries as a single metric
• Works bidirectionally (Datadog to OTel or OTel to Datadog)
• Preserves query aggregation semantics

Converting from Datadog to OTel

To include the equivalent OTel metrics in your query, wrap your Datadog query in equiv_otel:

equiv_otel(avg:nginx.net.connections)

This query:

1. Identifies the equivalent OTel metric (nginx.connections_current{state:active})
2. Combines timeseries from both metrics
3. Applies the aggregation (avg) across all datapoints

Converting from OTel to Datadog

The same works for including Datadog metrics in an OTel query:

equiv_otel(avg:nginx.connections_current{state:active})

The function works the same way in reverse, automatically including the equivalent Datadog metric (nginx.net.connections).

Further reading

Additional helpful documentation, links, and articles:

Metrics DocumentationDOCUMENTATION