apache kafka - Event sourcing in Flink -
i have flink application implemented following event-sourcing paradigm. both events , commands stored in several kafka topics.
the application has 2 startup modes: recovery , production. first, recovery mode used recover application state (a savepoint) events topics. in mode, commands not read @ all. once event topics have been processed, savepoint triggered manually (from command line) , application stopped. then, yarn process started in production mode. in mode application processes both events , commands.
i prefer execute process programatically. it, several questions arise... how can application itself:
- detect kafka sources have been readed?
- trigger savepoint programmatically?
- stop , start programmatically?
thank you!
Comments
Post a Comment