Follow the steps below to enable the new dependency map on the Service Page, and add inferred services to the Service Map and Service Catalog pages.

Inferred entities

Datadog can automatically discover the dependencies for an instrumented service, such as a database, a queue, or a third-party API even if that dependency hasn’t been instrumented yet. Datadog infers the presence of these dependencies and their performance based on information on the outbound requests of your instrumented services.

To determine the names and types of the inferred entities, Datadog uses span attributes. Inferred external APIs use the default naming scheme net.peer.name. For example, api.stripe.com, api.twilio.com, us6.api.mailchimp.com. Inferred databases use the default naming scheme db.instance.

If you’re using the Go, Java, NodeJS, PHP, .NET, or Ruby tracer, you can customize the default names for inferred entities.

Note: If you configure monitors, dashboards, or notebooks for a given inferred service during the beta, you may need to update them if the naming scheme changes.

Dependency map

Use the dependency map to visualize service-to-service communication and gain insight into system components such as databases, queues, and third-party dependencies. You can group dependencies by type and filter by Requests, Latency, or Errors to identify slow or failing connections.

Opt in

To opt in, you must adjust your Datadog Agent and APM Tracer configurations. Check the Global default service naming migration, to see if you need to take any migration actions.

Datadog Agent configuration

Requirements:

• Datadog Agent version >= 7.50.3.

Set the following in your datadog.yaml configuration file:

DD_APM_COMPUTE_STATS_BY_SPAN_KIND=true 
DD_APM_PEER_TAGS_AGGREGATION=true
DD_APM_PEER_TAGS='["_dd.base_service","amqp.destination","amqp.exchange","amqp.queue","aws.queue.name","bucketname","cassandra.cluster","cassandra.keyspace","db.cassandra.contact.points","db.couchbase.seed.nodes","db.hostname","db.instance","db.name","db.system","grpc.host","hazelcast.instance","hostname","http.host","messaging.destination","messaging.destination.name","messaging.kafka.bootstrap.servers","messaging.rabbitmq.exchange","messaging.system","mongodb.db","msmq.queue.path","net.peer.name","network.destination.name","peer.hostname","peer.service","queuename","rpc.service","rpc.system","server.address","streamname","tablename","topicname"]'

Helm

Include the same set of environment variables in your values.yaml file.

OpenTelemetry Collector

Minimum version recommended: opentelemetry-collector-contrib >= v0.95.0.

Example collector.yaml.

connectors:
  datadog/connector:
    traces:
      compute_stats_by_span_kind: true
      peer_tags_aggregation: true
      peer_tags: ["_dd.base_service","amqp.destination","amqp.exchange","amqp.queue","aws.queue.name","bucketname","cassandra.cluster","db.cassandra.contact.points","db.couchbase.seed.nodes","db.hostname","db.instance","db.name","db.system","grpc.host","hazelcast.instance","hostname","http.host","messaging.destination","messaging.destination.name","messaging.kafka.bootstrap.servers","messaging.rabbitmq.exchange","messaging.system","mongodb.db","msmq.queue.path","net.peer.name","network.destination.name","peer.hostname","peer.service","queuename","rpc.service","rpc.system","server.address","streamname","tablename","topicname"]

If your collector version is below v0.95.0, use an exporter configuration with the following peer_tags:

exporters:
  datadog:
    traces:
      compute_stats_by_span_kind: true
      peer_tags_aggregation: true
      peer_tags: ["_dd.base_service","amqp.destination","amqp.exchange","amqp.queue","aws.queue.name","bucketname","cassandra.cluster","db.cassandra.contact.points","db.couchbase.seed.nodes","db.hostname","db.instance","db.name","db.system","grpc.host","hazelcast.instance","hostname","http.host","messaging.destination","messaging.destination.name","messaging.kafka.bootstrap.servers","messaging.rabbitmq.exchange","messaging.system","mongodb.db","msmq.queue.path","net.peer.name","network.destination.name","peer.hostname","peer.service","queuename","rpc.service","rpc.system","server.address","streamname","tablename","topicname"]   

APM tracer configuration

Global default service naming migration

When you enable the DD_TRACE_REMOVE_INTEGRATION_SERVICE_NAMES_ENABLED environment variable, it improves how service-to-service connections and inferred services are represented in Datadog visualizations, across all supported tracing library languages and integrations.

Previously, some tracing libraries included the name of the associated integration in service name tagging. For example, .NET tagged gRPC calls as service:<DD_SERVICE>-grpc-client while Python tagged them as service:grpc-client. With this option enabled, all supported tracing libraries tag spans from the downstream services with the calling service’s name, service:<DD_SERVICE>, thereby providing a global default service name.

Consequently, if you have existing:

• APM metrics
• APM custom span metrics
• Trace analytics
• Retention filters
• Sensitive data scans
• Monitors, dashboards, or notebooks that query those things

Update those items to use the global default service tag (service:<DD_SERVICE>) instead.

Further reading

Additional helpful documentation, links, and articles:

Service ObservabilityDOCUMENTATION Sending Traces to DatadogDOCUMENTATION Add the Datadog Tracing LibraryDOCUMENTATION