The dogstatsd adapter is designed to deliver Istio metric instances to a listening DataDog agent.

This adapter supports the metric template.


Configuration parameter for the DataDog adapter. These params control how Mixer telemetry is transformed and sent to a dogstatsd agent.

The adapter assumes that a dogstatsd agent is running as a sidecar or at some other endpoint that the Mixer can reach. Any dimension that is a part of the metric is converted to a tag automatically. The configuration of the DataDog agent/daemon is outside the scope of the adapter.

Field Type Description
address string

Address of the dogstatsd server. Default: localhost:8125

prefix string

Prefix to prepend to all metrics handled by the adapter. Metric “bar” with prefix “foo.” becomes “” in DataDog. In order to make sure the metrics get populated into Datadog properly and avoid any billing issues, it’s important to leave the metric prefix to its default value of ‘istio.’ Default: “istio.”

bufferLength int32

Number of individual metrics to buffer before flushing metrics to the network. When buffered, metrics are flushed every 100ms or when the buffer is filled. When buffer is 0, metrics are not buffered. Default: 0

globalTags map<string, string>

Tags to add to every metric. “global”: “tag” becomes “global:tag” in DataDog Default: []

sampleRate double

Chance that any particular metric is sampled when emitted; can take the range [0, 1]. Default: 1

metrics map<string, Params.MetricInfo>

Map of a specific metric instance name -> info. If a metric’s instance name is not in the map then the metric will not be exported to DataDog.


Describes how to represent this metric in DataDog

Field Type Description
name string

Name of the metric in DataDog

type Params.MetricInfo.Type

The type of metric

tags map<string, string>

Tags to add to the metric in addition to the dimensions. “tag”: “val” becomes “tag:val” in DataDog Default: []


Describes the type of metric

Name Description

Default Unknown Type


Increments a DataDog counter


Sets the new value of a DataDog gauge


DISTRIBUTION is converted to a Timing Histogram for metrics with a time unit and a Histogram for all other units