There is one fetcher per broker, which is responsible for fetching messages in all consumed topics whose leader is on that broker. We have seen a fetcher being killed by a bug in Kafka. Also, if the broker is slow (e.g. due to I/O contention), the fetch rate could also be slower than expected.
Thanks, Jun On Tue, Dec 24, 2013 at 12:48 PM, Jason Rosenberg <j...@squareup.com> wrote: > Thanks Jun, > > I'm still not 100% clear on this. Is the min-fetch rate per topic, or > is it the lowest fetch rate over all topics? Or is it not topic > specific at all. > > What would cause this to slow down? Or is it more a measure of the > consumer's rate at processing messages and ability fetching new > messages (regardless of topic)? > > If we have some topics with no data at all, that match the consumer's > topic filter, should we expect to always see a min fetch rate of zero > (since at least one topic really has no messages)? > > Jason > > On Tue, Dec 24, 2013 at 1:10 PM, Jun Rao <jun...@gmail.com> wrote: > > I updated http://kafka.apache.org/documentation.html#monitoring > > > > Thanks, > > > > Jun > > > > > > On Mon, Dec 23, 2013 at 10:51 PM, Jason Rosenberg <j...@squareup.com> > wrote: > > > >> I'm realizing I'm not quite sure what the 'min fetch rate' metrics is > >> indicating, for consumers. Can someone offer an explanation? > >> > >> Is it related to the 'max lag' metric? > >> > >> Jason > >> >
