Azure Stream Analytics SU % Utilization keeps increasing -
i using stream analytics simple data pass-through scenario.
job has multiple sql server outputs (three) , 1 eventhub input. event count small. problem su % utilization keeps increasing. temporary solution restart job once or twice day.
am doing wrong? below sample of query. tried change queries use tumblingwindow, got same result.
select field_1, field_2, field_3, field_4, field_5 [out-alias-1] [in-alias] field_1 'event1:%' or field_1 'event2:%'; select field_1, field_3, field_6, field_7, field_8 [out-alias-2] [in-alias] field_1 'event3:%' or field_1 'event4:%'; select field_1, field_3, field_4, field_9, field_10 [out-alias-3] [in-alias] field_1 not 'event1:%' , field_1 not 'event2:%' , field_1 not 'event3:%' , field_1 not 'event4:%';
you may observe baseline su % utilization few or without input events, because system consumes amount of resource. amount of resource consumed system may fluctuate on time.
streaming units (sus) represent resources , power execute azure stream analytics job. sus provide way describe relative event processing capacity based on blended measure of cpu, memory, , read , write rates. each streaming unit corresponds 1mb/second of throughput. details, stream analytics key concepts.
the su % utilization metric indicator need scale azure stream analytics job. high su % utilization may result of large window in query, large events in input, large out of order tolerance window, or combination of above. partitioning query, or breaking down query more steps, , adding more sus scale tab both strategies avoid such condition. details, see scale azure stream analytics jobs.
Comments
Post a Comment