Prevent Incorrect Load-Aware Routing for AMD SGLang Backends (llm-d#2025)

weizhoublue · web-flow · commit 37fd7694033b · 2026-07-10T16:15:09.000-04:00
Signed-off-by: weizhoublue &lt;weizhou.lan@daocloud.io&gt;
diff --git a/guides/optimized-baseline/modelserver/amd/sglang/kustomization.yaml b/guides/optimized-baseline/modelserver/amd/sglang/kustomization.yaml
@@ -14,6 +14,7 @@ labels:
       llm-d.ai/model: Qwen3-32B
       llm-d.ai/accelerator-variant: gpu
       llm-d.ai/accelerator-vendor: amd
+      llm-d.ai/engine-type: sglang
     includeSelectors: true
     includeTemplates: true
 patches:
