Hi, we are looking at UMH-Core and Redpanda for a future product. Redpanda has some interesting features we like to use. To save on resources we would like to use UMH-Core's Redpanda. To prevent rouge micro services to publish unvalidated data on the 'umh.messages' topic we would like to enable ACL's. Are there any plans on the UMH-Core roadmap for this? Two things need to be done: enable sasl in the Redpanda config and have a way to run some rpk commands after the Redpanda starts.

at UHM we dont expose redpanda directly, as you land up with a lot of challenges that then each client needs to solve. and one bad client can break it for the rest. if you create a datamodel, we will autoamtically use the internal schema registry to enforce that nothing else lands up in the UNS and no wrong data. if you use our bridges, we will automatically properly use schema registry. however, external clients need to check themselves agsisnt schema registry, so one rogue esxternal client can all bypass it. you can create bridges that expose the data instead, e.g., HTTP API or push it somewhere else does that make sense?

Thanks for your answer. We are still exploring. Could you shed a light on what the challenges are each client needs to solve?

in kafka, the server is stupid and each client needs to check agsindt schema registry, handle backpressure, need to be aware of partitions, etc. etc. usually in manufacturing the clients are rather stupid and we want them to block them from doign stupid things (e.g., writing into wrong topics, etc.). so this is why we put the bridges in between as exactly that layer

Understood. Independent of the recommended architecture, does UMH-Core currently support or plan to support SASL and Redpanda ACLs?

we dont plan to, as you would not need it in our recommended architecture