Skip to main content
AI Gateway is OpenTelemetry (OTEL) compliant, making it easy to integrate with modern observability tools and platforms. Tracing and metrics can each be exported from the TrueFoundry dashboard to external OTEL backends. For Prometheus scraping of /metrics, labels, and the full metric catalog on self-hosted gateways, see Prometheus & Grafana integration.

​
Tracing

The OTEL Traces Exporter sends only trace data to your OTEL platform, not metrics. For OpenTelemetry metrics export, see Metrics below.
OpenTelemetry tracing allows you to capture detailed traces of requests as they flow through the AI Gateway. This enables debugging, performance analysis, and end-to-end visibility. Traces storage is managed by TrueFoundry. Optionally, traces can also be configured to be exported to any existing OTEL compatible observability platform. This will not affect the storage of traces managed by TrueFoundry.
OpenTelemetry traces storage and export flow in TrueFoundry

​
Export Traces to OTEL Compatible Observability Platforms

To export spans to your existing OTEL Platform:
OpenTelemetry Config form for exporting spans to existing OTEL platform request with spans
  1. Go to AI Gateway → Controls → Settings in the TrueFoundry dashboard
  2. Scroll down to the OTEL Config section and click the edit button
  3. Enable the OTEL Traces Exporter Configuration toggle
  4. Select either HTTP Configuration or gRPC Configuration based on your OTEL platform’s endpoint
  5. Enter the complete traces endpoint for eg., the traces endpoint for Grafana Cloud is https://otlp-gateway-prod-ap-south-1.grafana.net/otlp/v1/traces
  6. Select the encoding (only for HTTP Configuration) as either proto or json
  7. Optionally, enter the headers required to export spans to the OTEL provider for eg., authorization, region, etc..

​
Metrics

The OTEL Metrics Exporter sends only metrics to your OTEL platform, not traces. For OpenTelemetry traces export, see Tracing above.
OpenTelemetry metrics export lets you send gateway metrics to an OTLP-compatible observability platform. Configure this in the same OTEL Config editor as traces; it does not replace metrics stored and shown inside TrueFoundry.
OpenTelemetry metrics storage and export flow in TrueFoundry

​
Export Metrics to OTEL Compatible Observability Platforms

To export metrics to your existing OTEL platform:
OpenTelemetry Config form for exporting metrics to an OTEL platform
  1. Go to AI Gateway → Controls → Settings in the TrueFoundry dashboard
  2. Scroll down to the OTEL Config section and click the edit button
  3. Enable the OTEL Metrics Exporter Configuration toggle
  4. Select either HTTP Configuration or gRPC Configuration based on your OTEL platform’s metrics endpoint
  5. Enter the complete metrics endpoint—for example, Datadog uses https://otlp.datadoghq.com/v1/metrics
  6. Select the encoding (only for HTTP Configuration) as either proto or json
  7. Optionally, enter the headers required to export metrics to the OTEL provider—for example, API keys or vendor-specific headers
  8. Click Save

​
Example: Exporting Metrics to Datadog

Datadog supports OTLP metrics ingest natively. Use the following configuration: Step 1 — Set the environment variable on your tfy-llm-gateway deployment Datadog’s OTLP ingest endpoint only accepts delta-encoded metrics. Set this environment variable in your tfy-llm-gateway deployment before configuring the exporter: 
OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE=Delta
Step 2 — Configure the OTEL Metrics Exporter in TrueFoundry In the OTEL Config editor, enable OTEL Metrics Exporter Configuration and set:
FieldValue
ProtocolHTTP Configuration
Metrics Endpointhttps://otlp.datadoghq.com/v1/metrics
Encodingproto
Add the following headers:
HeaderValue
dd-api-key<your-datadog-api-key>
dd-otel-metric-config{"histograms":{"mode":"distributions"}}
The dd-otel-metric-config header enables the histogram translator, which converts OTLP histogram metrics into Datadog distribution metrics. This is required to query percentile aggregations such as p50:ai_gateway_request_processing_ms{*}, p95:ai_gateway_request_processing_ms{*}, and similar metrics in the Datadog Metrics Explorer.
For sites other than US1, use the region-specific endpoint—for example, https://otlp.datadoghq.eu/v1/metrics for the EU site. See the Datadog OTLP metrics ingest documentation for the full list of endpoints and additional dd-otel-metric-config options.
For scraping /metrics with Prometheus, OTLP endpoint patterns, and a full list of metric names and labels, see Prometheus & Grafana integration.