@jonas Flink's Fork-Join Pool drives only the actors, which are doing coordination. Unless your job is permanently failing/recovering, they don't do much.
On Thu, Jan 26, 2017 at 2:56 PM, Robert Metzger <rmetz...@apache.org> wrote: > Hi Jonas, > > The good news is that your job is completely parallelizable. So if you are > running it on a cluster, you can scale it at least to the number of Kafka > partitions you have (actually even further, because the Kafka consumers are > not the issue). > > I don't think that the scala (=akka) worker threads are really the thing > that slows everything done. These threads should usually idle. > I just tried it with Visualvm (I don't own a Jprofiler license :) ) and > you can nicely see what's eating up CPU resources in my job: > http://i.imgur.com/nqXeHdi.png > > > > > On Thu, Jan 26, 2017 at 1:23 PM, Jonas <jo...@huntun.de> wrote: > >> JProfiler >> >> >> >> -- >> View this message in context: http://apache-flink-user-maili >> ng-list-archive.2336050.n4.nabble.com/Improving-Flink-Per >> formance-tp11248p11311.html >> Sent from the Apache Flink User Mailing List archive. mailing list >> archive at Nabble.com. >> > >
