I agree that it does not matter too much :) (that why it's a nit)

However, I think that the maximum is inclusive, and only if we exceed it, we 
should force processing. From my understanding, "maximum idle time" is actually 
a lower bound (-> don't force processing until this time passed) because we 
cannot guarantee anyway to not exceed this threshold. I see your point why the 
name might be counter intuitive (even if I think the name is correct). If you 
interpret the name strictly, we would be allowed (or actually we would be 
required) to force processing before the time passed. This interpretation would 
make the parameter useless (ie, user tells us to idle max 5 minutes and we obey 
by forcing processing after 1 minute).

To me, the right interpretation is, "wait until this time passed and force 
processing asap if the time is exceeded". Chaning the name to 
`min.idle.time.ms` would be more precise, but I think it would be more 
confusing to users. 

[ Full content available at: https://github.com/apache/kafka/pull/5428 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to