For AI agents: A markdown version of this page is available at https://docs.datadoghq.com/opentelemetry/setup/otlp_ingest/serverless.md. A documentation index is available at /llms.txt.

OTLP Intake for Serverless

Overview

Send traces from serverless workloads directly to Datadog over HTTP/protobuf, without requiring a Datadog Agent or OpenTelemetry Collector. If your platform appears in the Managed platforms table, use its dedicated endpoint instead.

Serverless workloads can also send logs and metrics through the general OTLP logs and OTLP metrics intake endpoints. The resource attributes on this page apply to all signals your application exports.

Supported platforms:

  • AWS: Lambda, ECS Fargate
  • Azure: Container Apps, Web Apps (App Service), Azure Functions
  • GCP: Cloud Run, Cloud Run Functions, GKE Autopilot
Use direct ingest when running a Collector is impractical (for example, Lambda). If you can run a Collector, see OpenTelemetry Collector for metadata enrichment, normalization, and centralized sampling.

Prerequisites

The following configuration applies to all platforms.

Protocol: http/protobuf or http/json. grpc is not supported.

Required headers:

  • dd-api-key: Your Datadog API key.
  • dd-otlp-source: Set to serverless.
  • compute_stats: Set to true. Required for trace metrics.

Service name: Set OTEL_SERVICE_NAME to identify your service. Without it, traces appear as unknown_service.

Resource attributes: Set platform-specific attributes with OTEL_RESOURCE_ATTRIBUTES. See each cloud provider tab below for required and optional attributes.

export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT=""
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-service"

Setup

Based on your Datadog site, which is : Replace ${YOUR_ENDPOINT} with in the following examples.

Select your cloud provider for platform-specific resource attribute configuration:

    Lambda

    The AWS Distro for OpenTelemetry (ADOT) Lambda layer provides automatic instrumentation and resource detection for Lambda functions.

    Add the ADOT layer to your Lambda function and configure the following environment variables:

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-lambda-function"

    The ADOT layer handles resource attribute detection automatically. If you are not using ADOT, set resource attributes manually. cloud.provider is required. Set faas.id (a parseable Lambda ARN) for full platform identification; if faas.id is not available, set cloud.platform=aws_lambda instead:

    export OTEL_RESOURCE_ATTRIBUTES="cloud.provider=aws,faas.id=arn:aws:lambda:us-east-1:123456789012:function:my-function"
    AttributeRequiredDescription
    cloud.providerYesSet to aws
    faas.idRecommendedLambda function ARN (preferred for platform identification)
    cloud.platformConditionalSet to aws_lambda if faas.id is not set
    cloud.regionNoAWS region
    faas.nameNoFunction name
    faas.versionNoFunction version
    faas.instanceNoInstance identifier
    faas.max_memoryNoMax memory configured (bytes)
    aws.log.group.namesNoCloudWatch log group names (enables trace-log correlation)
    aws.log.stream.namesNoCloudWatch log stream names

    ECS Fargate

    ECS Fargate identification is driven by the task ARN and launch type, not by cloud.provider or cloud.platform. Configure the OpenTelemetry SDK to export traces directly from your ECS task:

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-ecs-service"
export OTEL_RESOURCE_ATTRIBUTES="aws.ecs.task.arn=arn:aws:ecs:us-east-1:123456789012:task/my-cluster/1234567890abcdef,aws.ecs.launchtype=fargate"
    AttributeRequiredDescription
    aws.ecs.task.arnYesECS task ARN
    aws.ecs.launchtypeYesSet to fargate (case-insensitive)
    cloud.providerNoDefaults to aws
    cloud.platformNoDefaults to aws_ecs
    cloud.regionNoAWS region
    cloud.availability_zoneNoAvailability zone
    aws.ecs.cluster.arnNoCluster ARN
    aws.ecs.task.familyNoTask definition family
    aws.ecs.task.idNoTask ID
    aws.ecs.task.revisionNoTask definition revision
    aws.log.group.namesNoCloudWatch log group names (enables trace-log correlation)
    aws.log.stream.namesNoCloudWatch log stream names
    The OpenTelemetry Collector's Azure resource detection processor only supports VMs. SDK-level Azure resource detectors support some platforms (Web Apps, Functions), but coverage varies by language SDK. Set cloud.provider, cloud.platform, and cloud.resource_id manually as the reliable path for all Azure serverless platforms.

    Container Apps

    Azure resource detector support for Container Apps varies by language SDK. Set resource attributes manually:

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-container-app"
export OTEL_RESOURCE_ATTRIBUTES="cloud.provider=azure,cloud.platform=azure.container_apps,cloud.resource_id=/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.App/containerApps/{appName}"

    Web Apps (App Service)

    Use the Azure resource detector SDK package (coverage varies by language SDK) or set OTEL_RESOURCE_ATTRIBUTES manually:

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-web-app"
export OTEL_RESOURCE_ATTRIBUTES="cloud.provider=azure,cloud.platform=azure.app_service,cloud.resource_id=/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.Web/sites/{appName}"

    Azure Functions

    Use the Azure resource detector SDK package (coverage varies by language SDK) or set OTEL_RESOURCE_ATTRIBUTES manually:

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-azure-function"
export OTEL_RESOURCE_ATTRIBUTES="cloud.provider=azure,cloud.platform=azure.functions,cloud.resource_id=/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.Web/sites/{functionAppName}"

    Resource attributes reference

    Platformcloud.providercloud.platformcloud.resource_id
    Container Appsazureazure.container_apps/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.App/containerApps/{appName}
    Web Appsazureazure.app_service/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.Web/sites/{appName}
    Azure Functionsazureazure.functions/subscriptions/{subscriptionId}/resourceGroups/{resourceGroup}/providers/Microsoft.Web/sites/{functionAppName}

    GCP resource detection works automatically with the GCP Resource Detector SDK package. Add it to your application dependencies to populate resource attributes without manual configuration.

    Cloud Run and Cloud Run Functions

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-cloud-run-service"

    The GCP Resource Detector SDK automatically populates: cloud.account.id, cloud.platform, cloud.provider, cloud.region, faas.id, faas.instance, faas.name, faas.version.

    GKE Autopilot

    export OTEL_EXPORTER_OTLP_TRACES_PROTOCOL="http/protobuf"
export OTEL_EXPORTER_OTLP_TRACES_ENDPOINT="${YOUR_ENDPOINT}"
export OTEL_EXPORTER_OTLP_TRACES_HEADERS="dd-api-key=${DD_API_KEY},dd-otlp-source=serverless,compute_stats=true"
export OTEL_SERVICE_NAME="my-gke-service"

    The GCP Resource Detector SDK automatically populates: cloud.account.id, cloud.platform, cloud.provider, cloud.region, host.id, k8s.cluster.name.

    Resource attributes reference

    PlatformAttributes populated by GCP Resource Detector
    Cloud Run / Cloud Run Functionscloud.account.id, cloud.platform, cloud.provider, cloud.region, faas.id, faas.instance, faas.name, faas.version
    GKE Autopilotcloud.account.id, cloud.platform, cloud.provider, cloud.region, host.id, k8s.cluster.name

    Further reading

    Additional helpful documentation, links, and articles:

    Datadog OTLP Intake EndpointDOCUMENTATION more more Datadog Serverless MonitoringDOCUMENTATION more more