I am developing an online prediction models for different customers. Customer data are saved in unique Cassandra DB. I need to initiate several streaming jobs for ach customer. Each streaming jobs using an oozie coordinator and a Spark SQL action.
I wanted to know if I can develop above system so actually to have several oozie coordinators in an oozie server (or equivalently several Falcon instances) for each customer?
UPDATE: Can I use an oozie bundle with several coordinators which can be configured dynamically at run-time?