Hi Everyone,

This is a very broad--perhaps a silly question, but I was wondering if a
small number of slow workers (low number of acks, low number tuples
executed) throttles overall throughput beyond the context of
max.spout.pending?.  For example, for a topology of 10 workers, would one
slow worker throttle the throughput of the other 9, again, beyond the
throttling provided by max.spout.pending?

Thanks

--John

Reply via email to