(OO) Support Tips: How the worker polling mechanism works
Each worker will hold a file called engine-webapp-xx.yy.jar, this file determines how often a worker will poll task from Central.
inBufferCapacity = "500"
numberOfExecutionThreads = "20"
maxDeltaBetweenDrains = "100"
The following configuration will work the following way:
Every 100 ms, it will retrieve up to 500 tasks and spread them across 20 threads.
Re: (OO) Support Tips: How the worker polling mechanism works
What exactly is the relationship between inBufferCapacity and the numberOfExecutionThreads? In this example, it was 500 buffer to 20 threads. If I had 200 execution threads, should I be configuring inBufferCapacity to 5000 so there is sufficient buffer?
Is there a calculation that I could use to ensure I wasn't starving my 200 threads of tasks?
And BTW, what exactly constitutes a "task" in the context of flow runs?