From 5b85d0efa423d5d0d3d0de424015f2d008d9fd14 Mon Sep 17 00:00:00 2001
From: joelnishanth <140015627+joelnishanth@users.noreply.github.com>
Date: Thu, 26 Mar 2026 13:22:07 -0500
Subject: [PATCH] discord: fix stale-socket reconnect crash from uncaught
 reconnect-exhausted error

---
 CHANGELOG.md                                  |  1 +
 .../src/monitor/provider.lifecycle.test.ts    | 45 +++++++++++++++++++
 .../discord/src/monitor/provider.lifecycle.ts | 17 ++++++-
 3 files changed, 62 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index ef0e4629b90..dda5b40980e 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -62,6 +62,7 @@ Docs: https://docs.openclaw.ai
 - Agents/failover: classify Codex accountId token extraction failures as auth errors so model fallback continues to the next configured candidate. (#55206) Thanks @cosmicnet.
 - Talk/macOS: stop direct system-voice failures from replaying system speech, use app-locale fallback for shared watchdog timing, and add regression coverage for the macOS fallback route and language-aware timeout policy. (#53511) thanks @hongsw.
 - Discord/gateway cleanup: keep late Carbon reconnect-exhausted errors suppressed through startup/dispose cleanup so Discord monitor shutdown no longer crashes on late gateway close events. (#55373) Thanks @Takhoffman.
+- Discord/gateway shutdown: treat expected reconnect-exhausted events during intentional lifecycle stop as clean shutdowns so startup-abort cleanup no longer surfaces false gateway failures. (#55324) Thanks @joelnishanth.
 
 ## 2026.3.24
 
diff --git a/extensions/discord/src/monitor/provider.lifecycle.test.ts b/extensions/discord/src/monitor/provider.lifecycle.test.ts
index a1cf1b8605f..9438843f95a 100644
--- a/extensions/discord/src/monitor/provider.lifecycle.test.ts
+++ b/extensions/discord/src/monitor/provider.lifecycle.test.ts
@@ -806,6 +806,51 @@ describe("runDiscordGatewayLifecycle", () => {
     }
   });
 
+  it("suppresses reconnect-exhausted as expected during intentional shutdown", async () => {
+    const { runDiscordGatewayLifecycle } = await import("./provider.lifecycle.js");
+    const pendingGatewayEvents: DiscordGatewayEvent[] = [];
+    const abortController = new AbortController();
+
+    const emitter = new EventEmitter();
+    const gateway = {
+      isConnected: true,
+      options: { reconnect: { maxAttempts: 50 } },
+      disconnect: vi.fn(),
+      connect: vi.fn(),
+      emitter,
+    };
+    getDiscordGatewayEmitterMock.mockReturnValueOnce(emitter);
+
+    const { lifecycleParams, runtimeLog, runtimeError } = createLifecycleHarness({
+      gateway,
+      pendingGatewayEvents,
+    });
+    lifecycleParams.abortSignal = abortController.signal;
+
+    // Start lifecycle; it yields at execApprovalsHandler.start(). We then
+    // queue a reconnect-exhausted event and abort. The lifecycle resumes,
+    // drains the event (with lifecycleStopping=true), and exits cleanly
+    // without reaching waitForDiscordGatewayStop.
+    const lifecyclePromise = runDiscordGatewayLifecycle(lifecycleParams);
+
+    pendingGatewayEvents.push(
+      createGatewayEvent(
+        "reconnect-exhausted",
+        "Max reconnect attempts (0) reached after code 1005",
+      ),
+    );
+    abortController.abort();
+
+    await expect(lifecyclePromise).resolves.toBeUndefined();
+
+    expect(runtimeLog).toHaveBeenCalledWith(
+      expect.stringContaining("ignoring expected reconnect-exhausted during shutdown"),
+    );
+    expect(runtimeError).not.toHaveBeenCalledWith(
+      expect.stringContaining("Max reconnect attempts"),
+    );
+  });
+
   it("does not push connected: true when abortSignal is already aborted", async () => {
     const { runDiscordGatewayLifecycle } = await import("./provider.lifecycle.js");
     const emitter = new EventEmitter();
diff --git a/extensions/discord/src/monitor/provider.lifecycle.ts b/extensions/discord/src/monitor/provider.lifecycle.ts
index c6cdd2d60f6..dd8ba184ec2 100644
--- a/extensions/discord/src/monitor/provider.lifecycle.ts
+++ b/extensions/discord/src/monitor/provider.lifecycle.ts
@@ -485,6 +485,15 @@ export async function runDiscordGatewayLifecycle(params: {
       );
       return "stop";
     }
+    // When we deliberately set maxAttempts=0 and disconnected (health-monitor
+    // stale-socket restart), Carbon fires "Max reconnect attempts (0)". This
+    // is expected — log at info instead of error to avoid false alarms.
+    if (lifecycleStopping && event.type === "reconnect-exhausted") {
+      params.runtime.log?.(
+        `discord: ignoring expected reconnect-exhausted during shutdown: ${event.message}`,
+      );
+      return "stop";
+    }
     params.runtime.error?.(danger(`discord gateway error: ${event.message}`));
     return event.shouldStopLifecycle ? "stop" : "continue";
   };
@@ -494,7 +503,13 @@ export async function runDiscordGatewayLifecycle(params: {
       if (decision !== "stop") {
         return "continue";
       }
-      if (event.type === "disallowed-intents") {
+      // Don't throw for expected shutdown events — intentional disconnect
+      // (reconnect-exhausted with maxAttempts=0) and disallowed-intents are
+      // both handled without crashing the provider.
+      if (
+        event.type === "disallowed-intents" ||
+        (lifecycleStopping && event.type === "reconnect-exhausted")
+      ) {
         return "stop";
       }
       throw event.err;