HandleHTTPRequest - NiFi

[#1]

Hey Guys,

General best practice question with this processor. When scheduling it to run I set this for a timer and zero seconds it will rack up ~550k tasks executed per 5min period on a cluster. If I limit the run schedule to every one second it greatly reduces this number and I have yet to see any failed packets sent to the endpoint.

Do I need to be concerned with packet loss if I limit run schedule to one second or is it recommended to let this run every zero seconds?

Thanks,

Zack

[#2]

@Zack Atkinson

The answer to your question will depend on the performance of your nifi install/cluster, the performance of your network, the Http Endpoint, as well as the configuration and setting of the downstream processors. By default 0 second should work and be equivalent to "always on".

Packet loss should not be a concern based on the one/zero second situation. Maybe that is a symptom of hitting the Http endpoint faster than it can handle?

Finding the right balance that works for your Use Case is going to be some combination of all the above in which case tuning the processor for every 1 sec, 5, sec, etc may be necessary.