From cf1ec7b1c6fc37914a0a48d8776a878d2b58a146 Mon Sep 17 00:00:00 2001
From: Jerry Yan <792602257@qq.com>
Date: Wed, 6 May 2026 11:40:34 +0800
Subject: [PATCH] fix(app): give dispatcher.Shutdown its own ctx in ctx.Done
 branch

Final cross-cutting review finding: shutCtx was shared between
server.Shutdown and dispatcher.Shutdown, with jobs.Stop's wall-clock
time burning the budget in between. Under load, dispatcher.Shutdown
could receive an already-dead ctx and skip draining in-flight async
notify dispatches (including webhook enqueues). Mirror the errCh
branch by creating a fresh 10s drainCtx for dispatcher.Shutdown.
---
 internal/app/app.go | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/internal/app/app.go b/internal/app/app.go
index 32f4b63..b100522 100644
--- a/internal/app/app.go
+++ b/internal/app/app.go
@@ -418,8 +418,13 @@ func (a *App) Run(ctx context.Context) error {
 			a.jobs.Stop(stopCtx)
 			stopCancel()
 		}
-		// server 已停止接收新请求；等异步通知投递结束再关 pool
-		if err := a.dispatcher.Shutdown(shutCtx); err != nil {
+		// server 已停止接收新请求；等异步通知投递结束再关 pool。
+		// 用独立的 drainCtx：shutCtx 的 30s 预算已被 server.Shutdown 与
+		// jobs.Stop 的实际墙钟时间消耗，复用可能让 dispatcher 拿到已死 ctx
+		// 而立即返回，丢失在途的 webhook enqueue（DB 写）。
+		drainCtx, drainCancel := context.WithTimeout(context.Background(), 10*time.Second)
+		defer drainCancel()
+		if err := a.dispatcher.Shutdown(drainCtx); err != nil {
 			a.log.Warn("notify dispatcher drain", "err", err.Error())
 		}
 		a.pool.Close()