Azure Stream Analytics不会输出所有输入事件 [英] Azure Stream Analytics not outputting all input events

问题描述

我们有一个系统，该系统通过Open Automation软件将标签数据发送到Azure IOT HUB实例，然后再发送到Azure Stream Analytics，再输出到Azure SQL DW.

We have a system that sends tag data through Open Automation Software to an Azure IOT HUB instance and then on to Azure Stream Analytics that then outputs to an Azure SQL DW.

系统正常运行(即– SA输出所有输入事件)，每30秒最多可容纳26个标签值的数据(数据事件很小，.16 KB).

The system works correctly (ie – SA outputs all input events) up to 26 tags worth of data (data events are small, .16 KB) every 30 seconds.

从测试运行的一开始，在价值27个标签的数据SA上的输出开始失败.接收到来自IOT HUB的所有输入事件，但是SA仅将这些事件的子集输出到SQL DW. SU％利用率仅为10，因此 有足够的可用内存.从调查来看，显然没有能力对SA输出进行分区，因为目标是由Azure自动管理的目标.

At 27 tags worth of data SA begins to fail on outputs from the very beginning of the test run. All of the input events from IOT HUB are received, however only a subset of those events are output by SA to the SQL DW. SU % Utilization is only 10, so there is plenty of memory available. From investigation then there’s apparently no ability to partition the SA output as the target is one that is automatically managed by Azure.

SA作业图确实显示了有时积压的输入事件，但是该计数与输出事件数之间的差值相差不大.活动日志显示无写入或其他错误.

The SA job diagram does show input events backlogged upon occasion, but the delta between that count and the number of output events is nowhere near the same.  The Activity Log shows no Write or other errors.

我们缺少什么可以纠正此问题并允许平台正确缩放?

What are we missing that could correct this and allow the platform to scale correctly?

推荐答案

我已经从信息中与我们的工程师进行了验证由您提供，这似乎不是ASA问题.问题可能是由于查询逻辑或SQL目标架构

I have verified with our engineers, from the information provided by you this doesn't seem like ASA issue. The issue might be due to the query logic or SQL destination schema

这篇关于Azure Stream Analytics不会输出所有输入事件的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！