Just implemented a Queue benchmark (https://testqueueperforman.mxapps.io/) between the Queue and Process Queue modules to see whether there is a performance difference in terms of queue overhead. The queue configurations are the same (5 threads and thread affinity/priority of 5) and they perform the same tasks in a Mendix Free Cloud app container. 1x microflow to create a new benchmark object 10.000x empty microflow 1x microflow to set the duration of the benchmark See the results below for a 5-pass benchmark, the new Queue provides a factor 7.4 performance improvement compared to the Process Queue in terms of queue overhead. If you use the queue for long running jobs, queue overhead becomes less important. Do you like to try it yourself? Open the Queue Performance Benchmark. I did not investigate the cause of the difference in performance yet, but I think this requires some deeper analysis of the Java source code that is executed by both queue modules. Does anyone have suggestions?
asked
Menno de Haas
1 answers
0
Thanks for all your feedback by email. New version has been released and module has been downloaded 23 times last month!