feat: add guide for trace support (#2212)

The metrics doc is already very huge and long, keep the trace in another
doc as trace details will go long in the future.

Author: JeffLuoo

mkdocs.yml

@@ -73,7 +73,9 @@ nav:
       - Rollout:
         - Adapter Rollout: guides/adapter-rollout.md
         - InferencePool Rollout: guides/inferencepool-rollout.md
-      - Metrics and Observability: guides/metrics-and-observability.md
+      - Observability:
+        - Metrics: guides/metrics-and-observability.md
+        - Traces: guides/trace.md
       - Configuration Guide:
           - Configuring the EndPoint Picker via configuration YAML file: guides/epp-configuration/config-text.md
           - Prefix Cache Aware Plugin: guides/epp-configuration/prefix-aware.md
@@ -85,7 +87,7 @@ nav:
       - Conformance Tests: guides/conformance-tests.md
   - Performance:
     - Benchmark: performance/benchmark/index.md
-    - Advanced Benchmarking Configs: 
+    - Advanced Benchmarking Configs:
       - Prefix Cache Aware: performance/benchmark/advanced-configs/prefix-cache-aware.md
       - Decode Heavy Workload: performance/benchmark/advanced-configs/decode-heavy-workload.md
       - Prefill Heavy Workload: performance/benchmark/advanced-configs/prefill-heavy-workload.md

site-src/guides/trace.md

@@ -0,0 +1,40 @@
+# Trace
+
+This guide describes the current state of trace and how to use them.
+
+## Requirements
+
+You would need to enable the trace feature with following helm option:
+
+```yaml
+inferenceExtension:
+  tracing:
+    enabled: true
+    otelExporterEndpoint: "http://localhost:4317"
+    sampling:
+      sampler: "parentbased_traceidratio"
+      samplerArg: "0.1"
+```
+
+- `otelExporterEndpoint`: Points to your OpenTelemetry collector endpoint.
+- `sampler`: Currently, only `parentbased_traceidratio` is supported. This sampler respects the parent span's sampling decision and applies the configured ratio for root spans.
+- `samplerArg`: Base sampling rate for new traces, range [0.0, 1.0]. For example, "0.1" enables 10% sampling.
+
+## Span Coverage
+
+Currently, the inference gateway covers the entry point of the external processing request.
+
+- **Tracer Name**: `gateway-api-inference-extension`
+- **Span Name**: `gateway.request`
+
+This span is the root span the entire lifecycle of an external processing request from Envoy, including header and body processing, scheduling decisions, and response handling.
+
+## Attributes
+
+### Span Attributes
+
+Currently, the `gateway.request` span does not include custom attributes. It primarily serves as a container for the request's execution time and provides context for child spans in downstream services.
+
+## Context Propagation
+
+The inference gateway supports distributed tracing by propagating the trace context to downstream services (e.g., model servers).