how to achieve parallelism in Azure Stream Analytics -


if understanding right, 1 stream analytic jobs's instance runs @ time. , after finishes current set of events, next set of events pulled event hub. if sequential.

if processing takes 20 milliseconds, other events have wait many milliseconds. wondering if sequential operation sufficient in production load?

i aware of partitionby clause, since using iot hub, cannot use partitionid/ partitionkey.

thanks in advance

all messages same deviceid sent same partitionid. if query ever looks @ 1 deviceid @ time, can still use partitionid , process each partition independently. examples of queries select, filter queries , aggregates include deviceid in key.

if queries @ multiple deviceids @ time (for example, counting total number of messages in window), have 2 options. can partial aggregates first in parallel , combine them global aggregate. or use query without partition by.

also, azure stream analytics not messages 1 one incur kind of delays mentioned in question.


Comments

Popular posts from this blog

angular - Ionic slides - dynamically add slides before and after -

minify - Minimizing css files -

Add a dynamic header in angular 2 http provider -