Use this guide when an instrumented application runs in a Kubernetes workload and exports OTLP to an in-cluster Collector or trace proxy.

Select a reachable endpoint

The product setup documentation lists these default trace-proxy service endpoints for Kubernetes:

ProtocolEndpoint
OTLP/gRPCopsramp-tracing-proxy.opsramp-tracing-proxy.svc.cluster.local:9090
OTLP/HTTPhttp://opsramp-tracing-proxy.opsramp-tracing-proxy.svc.cluster.local:8082

These are defaults, not a guarantee for every cluster. Confirm the service name, namespace, listener ports, TLS termination, protocol, and authentication settings in the tenant’s Trace Setup instructions. If a Collector is the application-facing receiver, use that Collector’s service endpoint instead.

Configure the workload

Set the OpenTelemetry SDK’s service name, OTLP endpoint, and protocol through the workload configuration. For example, an HTTP/protobuf exporter can use:

env:
  - name: OTEL_SERVICE_NAME
    value: orders-api
  - name: OTEL_EXPORTER_OTLP_ENDPOINT
    value: http://opsramp-tracing-proxy.opsramp-tracing-proxy.svc.cluster.local:8082
  - name: OTEL_EXPORTER_OTLP_PROTOCOL
    value: http/protobuf

Use the SDK’s supported settings and the endpoint assigned to your cluster. Configure required authentication separately using the tenant’s instructions.

Protect credentials

Store secrets in a Kubernetes Secret or your organization’s approved external secret provider. Reference the secret from the workload or Collector without committing its value to manifests or source control. Restrict access to the namespace and secret, then roll out the workload change.

Verify

Check that the pod is ready, the application can resolve and reach the selected endpoint, and exporter logs show successful exports. Then follow Verify Trace Ingestion.