That was the initial approach, but creating & closing producer for each message kills throughput
hence the idea to start it once and reuse
you don’t need to create so many loggers; just use one for that one kafka instance, and disable the events you don’t need
you can configure idle_timeout, connection_limit and max_deliveries_per_connection for that queue using your shaping configuration and manage the lifecycle that way
Alright, will try this, thanks!