I am currently trying to run our Spark application with a REST API on Kubernetes, using the Spark Operator.
I got the driver and executors up and running, so that's all working. The Spark application is using a Gunicorn, to provide access to it. This gunicorn runs on a random port. This port, however, seems unknown to Kubernetes itself or to the spark operator, or the driver.
Our Spark/Yarn setup makes use of Zookeeper to register the ip and port, but if possible, I'd be nice to have this functionality with Kubernetes itself.
Is there any way to make this port known to Kubernetes, so it can be served with a Kubernetes service object?
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
I am currently trying to run our Spark application with a REST API on Kubernetes, using the Spark Operator.
I got the driver and executors up and running, so that's all working. The Spark application is using a Gunicorn, to provide access to it. This gunicorn runs on a random port. This port, however, seems unknown to Kubernetes itself or to the spark operator, or the driver.
Our Spark/Yarn setup makes use of Zookeeper to register the ip and port, but if possible, I'd be nice to have this functionality with Kubernetes itself. Is there any way to make this port known to Kubernetes, so it can be served with a Kubernetes service object?