Similar thing for SQL jobs, can't find an answer whether its possible but if its possible for CEP, it should be possible for others as well

Looks like there is an issue for CEP, but not getting much action https://issues.apache.org/jira/plugins/servlet/mobile#issue/FLINK-7129

I think their approach is deatiled here https://getindata.com/blog/dynamic-sql-processing-with-apache-flink/ But not sure about the approach though

I found this git repo: https://github.com/getindata/flink-dynamic-cep-demo Looks like they have a Flink Job with a function that spins up a minicluster for every rule they receive. Have mixed feelings on this approach, afaik the minicluster is intended for testing purposes not production use - and spinning up a minicluster for each ruke might become expensive