fix(queue): 自动救援失活 Worker 遗留任务

新增基于任务状态、执行租约和 Worker 心跳的孤儿 River Job 精确识别，每 15 秒将确认失活的 Job 恢复为可重试，避免等待 River 一小时通用救援窗口。运行时恢复同步清除中断及终态任务的 admission、执行令牌和并发租约，防止状态残留污染队列指标。正常运行中的长任务仍由 running 状态和活跃心跳保护。验证通过完整 Go 测试、go vet、竞态检查、迁移安全检查和 PostgreSQL 18 集成测试。
2026-07-29 23:44:29 +08:00
parent 98820378b7
commit f3dd7cd262
6 changed files with 369 additions and 3 deletions
@@ -47,13 +47,14 @@ func main() {
 	if recovery, err := db.RecoverInterruptedRuntimeState(ctx); err != nil {
 		logger.Error("recover interrupted runtime state failed", "error", err)
 		os.Exit(1)
-	} else if recovery.ReleasedConcurrencyLeases > 0 || recovery.ReleasedRateReservations > 0 || recovery.FailedAttempts > 0 || recovery.FailedTasks > 0 || recovery.RequeuedAsyncTasks > 0 {
+	} else if recovery.ReleasedConcurrencyLeases > 0 || recovery.ReleasedRateReservations > 0 || recovery.FailedAttempts > 0 || recovery.FailedTasks > 0 || recovery.RequeuedAsyncTasks > 0 || recovery.CleanedTaskAdmissions > 0 {
 		logger.Warn("interrupted runtime state recovered",
 			"releasedConcurrencyLeases", recovery.ReleasedConcurrencyLeases,
 			"releasedRateReservations", recovery.ReleasedRateReservations,
 			"failedAttempts", recovery.FailedAttempts,
 			"failedTasks", recovery.FailedTasks,
 			"requeuedAsyncTasks", recovery.RequeuedAsyncTasks,
+			"cleanedTaskAdmissions", recovery.CleanedTaskAdmissions,
 		)
 	}