Use a kafka api in python run uc
Use a kafka api in the python run uc script to avoid using the kafka client deployed in kubernetes
Activity
added modul/theodolite label
Seems to be hard to implement, when we want to use the script also from outside of Kubernetes. The api need a bootstrap server to connect to a Kafka broker. Then it discovers the other servers through this. But the Kafka brokers are headless services and are not easy to access from outside of Kafka.
Yes, that's a good point. I suggest to encapsulate the calls to other services such as Kafka, Zookeeper, Prometheus, and maybe others, such that when we can an implementation using the current kubectl approach and an implementation calling the corresponding services directly.
Prospectively, running Theodolite in Kubernetes jobs seems definitely to be the way to go, but till that, I think, there's still a lot to do.
Edited by Sören Henning-
I'm closing this issue in favor of #126 (closed)