I'm trying to run multiple independent CEP patterns. They're basic patterns, just one input followed by another and my flink job runs fine when just using 1 pattern. If i try to scale this up to add multiple CEP patterns, 200 for example, I start getting memory errors on my cluster. I can definitely add more memory, but I want to know if there is an accepted way to run multiple patterns.
Currently I am just doing a flatselect on the output of CEP.pattern and then making a list of all these inputs and putting a sink on all of the streams. What this is doing is creating a different stream for each pattern, which is causing the output of the Logical Plan in the UI to be too big to even see. Does anybody know of any better way to do this? -- View this message in context: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-CEP-Patterns-tp12871.html Sent from the Apache Flink User Mailing List archive. mailing list archive at Nabble.com.
