From 1c7881ec09893dc57f4e02aae3ead64ec932bca0 Mon Sep 17 00:00:00 2001
From: wenghuayang863 <381421746@qq.com>
Date: Mon, 11 May 2026 00:43:58 +0800
Subject: [PATCH] fix(runtime-fallback): match Volcano Engine 'exceeded the
 usage quota' errors

Volcano Engine sends quota exceeded errors with the words in reverse
order: 'You have exceeded the 5-hour usage quota'. The existing
patterns required 'quota' to precede 'exceeded', so they never matched.

- Add /exceeded.*quota/i and /usage.?quota/i to RETRYABLE_ERROR_PATTERNS
- Add exceeded.*quota and usage\s*quota to AUTO_RETRY_PATTERNS
- Add regression tests for both detection paths

Fixes: runtime-fallback not triggering on Volcano Engine quota errors
---
 .../auto-retry-signal.test.ts                 | 41 +++++++++++++++++++
 .../runtime-fallback/auto-retry-signal.ts     |  2 +-
 src/hooks/runtime-fallback/constants.ts       |  2 +
 .../quota-error-classifier.regression.test.ts | 15 +++++++
 4 files changed, 59 insertions(+), 1 deletion(-)
 create mode 100644 src/hooks/runtime-fallback/auto-retry-signal.test.ts

diff --git a/src/hooks/runtime-fallback/auto-retry-signal.test.ts b/src/hooks/runtime-fallback/auto-retry-signal.test.ts
new file mode 100644
index 000000000..e734b309b
--- /dev/null
+++ b/src/hooks/runtime-fallback/auto-retry-signal.test.ts
@@ -0,0 +1,41 @@
+import { describe, expect, test } from "bun:test"
+
+import { extractAutoRetrySignal } from "./auto-retry-signal"
+
+describe("extractAutoRetrySignal", () => {
+  test("detects Volcano Engine 'exceeded the usage quota' signal", () => {
+    //#given
+    const info = {
+      status: "You have exceeded the 5-hour usage quota. It will reset at 2026-05-11 01:20:12 +0800 CST.",
+    }
+
+    //#when
+    const signal = extractAutoRetrySignal(info)
+
+    //#then
+    expect(signal).toBeDefined()
+    expect(signal?.signal).toContain("exceeded")
+  })
+
+  test("detects standard 'quota exceeded' signal", () => {
+    //#given
+    const info = { message: "Quota exceeded for model gpt-4" }
+
+    //#when
+    const signal = extractAutoRetrySignal(info)
+
+    //#then
+    expect(signal).toBeDefined()
+  })
+
+  test("returns undefined for non-retryable info", () => {
+    //#given
+    const info = { message: "Something went wrong" }
+
+    //#when
+    const signal = extractAutoRetrySignal(info)
+
+    //#then
+    expect(signal).toBeUndefined()
+  })
+})
diff --git a/src/hooks/runtime-fallback/auto-retry-signal.ts b/src/hooks/runtime-fallback/auto-retry-signal.ts
index 1d33edbee..9e2e9ab67 100644
--- a/src/hooks/runtime-fallback/auto-retry-signal.ts
+++ b/src/hooks/runtime-fallback/auto-retry-signal.ts
@@ -5,7 +5,7 @@ export interface AutoRetrySignal {
 const AUTO_RETRY_PATTERNS: Array<(combined: string) => boolean> = [
   (combined) => /retrying\s+in/i.test(combined),
   (combined) =>
-    /(?:too\s+many\s+requests|quota\s+will\s+reset\s+after|quota\s*exceeded|usage\s+limit|rate\s+limit|limit\s+reached|all\s+credentials\s+for\s+model|cool(?:ing)?\s*down|exhausted\s+your\s+capacity)/i.test(combined),
+    /(?:too\s+many\s+requests|quota\s+will\s+reset\s+after|quota\s*exceeded|exceeded.*quota|usage\s+limit|usage\s*quota|rate\s+limit|limit\s+reached|all\s+credentials\s+for\s+model|cool(?:ing)?\s*down|exhausted\s+your\s+capacity)/i.test(combined),
 ]
 
 export function extractAutoRetrySignal(info: Record<string, unknown> | undefined): AutoRetrySignal | undefined {
diff --git a/src/hooks/runtime-fallback/constants.ts b/src/hooks/runtime-fallback/constants.ts
index 19a7cad56..e835b4099 100644
--- a/src/hooks/runtime-fallback/constants.ts
+++ b/src/hooks/runtime-fallback/constants.ts
@@ -27,6 +27,8 @@ export const RETRYABLE_ERROR_PATTERNS = [
   /too.?many.?requests/i,
   /quota\s+will\s+reset\s+after/i,
   /quota.?exceeded/i,
+  /exceeded.*quota/i,
+  /usage.?quota/i,
   /exhausted\s+your\s+capacity/i,
   /all\s+credentials\s+for\s+model/i,
   /cool(?:ing)?\s+down/i,
diff --git a/src/hooks/runtime-fallback/quota-error-classifier.regression.test.ts b/src/hooks/runtime-fallback/quota-error-classifier.regression.test.ts
index 1979ddc30..737db208c 100644
--- a/src/hooks/runtime-fallback/quota-error-classifier.regression.test.ts
+++ b/src/hooks/runtime-fallback/quota-error-classifier.regression.test.ts
@@ -56,4 +56,19 @@ describe("runtime-fallback quota error regressions", () => {
     // quota errors trigger fallback to next configured model
     expect(retryable).toBe(true)
   })
+
+  test("classifies Volcano Engine 'exceeded the usage quota' as retryable", () => {
+    //#given
+    const error = {
+      name: "SessionRetry",
+      message: "You have exceeded the 5-hour usage quota. It will reset at 2026-05-11 01:20:12 +0800 CST. We recommend using a different model.",
+    }
+
+    //#when
+    const retryable = isRetryableError(error, [429, 500, 502, 503, 504])
+
+    //#then
+    // Volcano Engine quota errors trigger fallback to the next model
+    expect(retryable).toBe(true)
+  })
 })