feat: increase timeout for inferencepool extproc (envoyproxy#1251)

Xunzhuo · web-flow · commit 0649d877a413 · 2025-09-30T21:46:08.000+08:00
**Description**

This PR increased timeout from 5s for inferencepool extproc, preventing
timeout from LLM

Signed-off-by: bitliu &lt;bitliu@tencent.com&gt;
diff --git a/internal/extensionserver/inferencepool.go b/internal/extensionserver/inferencepool.go
@@ -310,7 +310,7 @@ func buildHTTPFilterForInferencePool(pool *gwaiev1a2.InferencePool) *extprocv3.E
 			ResponseTrailerMode: extprocv3.ProcessingMode_SEND,
 		},
 		AllowModeOverride: allowModeOverride,
-		MessageTimeout:    durationpb.New(5 * time.Second),
+		MessageTimeout:    durationpb.New(300 * time.Second),
 		FailureModeAllow:  false,
 	}
 }

Original file line number	Diff line number	Diff line change
`@@ -310,7 +310,7 @@ func buildHTTPFilterForInferencePool(pool gwaiev1a2.InferencePool) extprocv3.E`
`310`	`310`	`ResponseTrailerMode: extprocv3.ProcessingMode_SEND,`
`311`	`311`	`},`
`312`	`312`	`AllowModeOverride: allowModeOverride,`
`313`		`- MessageTimeout: durationpb.New(5 * time.Second),`
	`313`	`+ MessageTimeout: durationpb.New(300 * time.Second),`
`314`	`314`	`FailureModeAllow: false,`
`315`	`315`	`}`
`316`	`316`	`}`