Post-upgrade to 23.4 we experienced these issues.
The delay in profile processing is no longer an issue.
We identified the root cause for the delay in the profile deployment in some scenarios and delivered a fix in the configuration_reader_service probe, and now all of the profiles are processed fine.
configuration_reader_service-23.4.1-20240223.153110-19.zip
Later a defect was identified in the discovery server's nis_cache processing and a fix was delivered and now all robot_nis_cache queues were processed without any issues. Then we started seeing the metrics for which QoS is available in the DB.
discovery_server_23.40T2.zip