节流:仅在队列触发的Azure功能完成后才删除队列消息吗?

剧透:似乎没有，所以我想知道如何做到这一点.

Spoiler: It doesn't seem to, so I want to know how to accomplish this.

在这种情况下，我需要进行大量工作(就我而言，是图像处理并向数据库报告)，我想使用Azure函数进行处理.问题是，如果我产生太多实例，我的Azure SQL数据库很快就会不知所措 并超时.因此，我需要做的是限制Azure Function的并发实例的数量.

I have a situation in which I have a large batch of work (in my case, image processing and reporting to a database) I want processed using an Azure Function.  The problem is that if I spawn too many instances, my Azure SQL Database quickly gets overwhelmed and times out.  So what I need to do is throttle the number of concurrent instances of the Azure Function.

我已经尝试使用Queue Storage + Queue Trigger作为该功能.我将图像上传到Blob存储，然后将消息添加到队列中，描述此上传，该消息成功触发了Azure功能.但是尽管有各种批处理大小 /newBatchThreshold/maxDequeueCount设置我尝试过，似乎仍然会同时产生太多的函数实例，并且Azure SQL数据库不可避免地会超时.

I've tried making use of Queue Storage + Queue Trigger for the function.  I upload an image to blob storage, then add a message to a Queue describing this upload which successfully triggers the Azure Function.  But despite the various batchSize / newBatchThreshold / maxDequeueCount settings I've tried, it still seems that too many function instances spawn at once and the Azure SQL database calls inevitably time out.

所以我的问题之一是:是否仅在函数实例完成后才从队列中删除消息，还是只是将消息从队列中拉出并且函数尽快运行?我正试图以一种可靠的方式来减慢速度， 但我仍然不知道如何做到这一点.

So one of my questions is: Does the message removal from the queue happen only after a function instance completes, or do messages just get pulled off the queue and functions run as quickly as possible?  I'm trying to slow this down in a reliable way, but I still can't tell how to accomplish this.