Has anyone ever used onyx with in memory zookeeper in production for a simple system?

I am envisioning an application that submits hourly spark jobs to a YARN cluster. The application would use juxt/joplin to migrate target databases when deployed, as well as killing currently running jobs. Onyx would simply handle the schedule of jobs, the UI, retries, and resolving the DAG of spark jobs.

Or perhaps there is a simpler system I could leverage.