fix: E11000 duplicate key error when creating timeout retry task

[sliss]

Problem

When a task times out and has retryOnTimeoutCount > 0, the expireTimedOutTasks() function creates a retry task by spreading task.toObject({ virtuals: false }). This includes the _id field from the original task, causing MongoDB to throw an E11000 duplicate key error when trying to insert the retry task.

MongoServerError: E11000 duplicate key error collection: db.tasks index: _id_ dup key: { _id: ObjectId('...') }

Solution

1. Copy the task data and delete the _id field before creating the retry task, allowing MongoDB to generate a new _id
2. Decrement retryOnTimeoutCount so retries are properly limited (without this, the retry task would have the same count and retry indefinitely)
3. Reset execution state fields (startedRunningAt, error, result, etc.) so the retry task starts fresh

Changes

// Before (buggy)
if (task.retryOnTimeoutCount > 0) {
  await Task.create({
    ...task.toObject({ virtuals: false }),
    status: 'pending',
    schedulingTimeoutAt: now.valueOf() + 10 * 60 * 1000
  });
}

// After (fixed)
if (task.retryOnTimeoutCount > 0) {
  const taskData = task.toObject({ virtuals: false });
  delete taskData._id;
  await Task.create({
    ...taskData,
    status: 'pending',
    retryOnTimeoutCount: task.retryOnTimeoutCount - 1,
    startedRunningAt: null,
    finishedRunningAt: null,
    workerName: null,
    error: null,
    result: null,
    timeoutAt: null,
    schedulingTimeoutAt: now.valueOf() + 10 * 60 * 1000
  });
}