fix(infra): treat undici fetch failed as transient unhandled rejection

CHANGELOG.md

@@ -85,6 +85,7 @@ Docs: https://docs.openclaw.ai
 - Plugins/CLI: make `openclaw plugins enable` and plugin install/link flows update allowlists via shared plugin-enable policy so enabled plugins are not left disabled by allowlist mismatch. (#23190) Thanks @downwind7clawd-ctrl.
 - Memory/QMD: add optional `memory.qmd.mcporter` search routing so QMD `query/search/vsearch` can run through mcporter keep-alive flows (including multi-collection paths) to reduce cold starts, while keeping searches on agent-scoped QMD state for consistent recall. (#19617) Thanks @nicole-luxe and @vignesh07.
 - Chat/UI: strip inline reply/audio directive tags (`[[reply_to_current]]`, `[[reply_to:<id>]]`, `[[audio_as_voice]]`) from displayed chat history, live chat event output, and session preview snippets so control tags no longer leak into user-visible surfaces.
+- Infra/Network: classify undici `TypeError: fetch failed` as transient in unhandled-rejection detection even when nested causes are unclassified, preventing avoidable gateway crash loops on flaky networks. (#14345) Thanks @Unayung.
 - Telegram/Retry: classify undici `TypeError: fetch failed` as recoverable in both polling and send retry paths so transient fetch failures no longer fail fast. (#16699) thanks @Glucksberg.
 - BlueBubbles/DM history: restore DM backfill context with account-scoped rolling history, bounded backfill retries, and safer history payload limits. (#20302) Thanks @Ryan-Haines.
 - BlueBubbles/Private API cache: treat unknown (`null`) private-API cache status as disabled for send/attachment/reply flows to avoid stale-cache 500s, and log a warning when reply/effect features are requested while capability is unknown. (#23459) Thanks @echoVic.

src/infra/unhandled-rejections.test.ts

@@ -79,6 +79,12 @@ describe("isTransientNetworkError", () => {
     expect(isTransientNetworkError(error)).toBe(true);
   });
 
+  it("returns true for fetch failed with unclassified cause", () => {
+    const cause = Object.assign(new Error("unknown socket state"), { code: "UNKNOWN" });
+    const error = Object.assign(new TypeError("fetch failed"), { cause });
+    expect(isTransientNetworkError(error)).toBe(true);
+  });
+
   it("returns true for nested cause chain with network error", () => {
     const innerCause = Object.assign(new Error("connection reset"), { code: "ECONNRESET" });
     const outerCause = Object.assign(new Error("wrapper"), { cause: innerCause });

src/infra/unhandled-rejections.ts

@@ -94,12 +94,10 @@ export function isTransientNetworkError(err: unknown): boolean {
     return true;
   }
 
-  // "fetch failed" TypeError from undici (Node's native fetch)
+  // "fetch failed" TypeError from undici (Node's native fetch).
+  // Treat as transient regardless of nested cause code because causes vary
+  // across runtimes and can be unclassified even for real network faults.
   if (err instanceof TypeError && err.message === "fetch failed") {
-    const cause = getErrorCause(err);
-    if (cause) {
-      return isTransientNetworkError(cause);
-    }
     return true;
   }