This message was deleted.

..

Indeed, it was designed precisely to indicate the health of the broker. Currently, there is a known issue where the health response may also fail if the bookie cluster is unavailable. Nonetheless, this issue is not particularly severe, as you can preemptively address bookie storage problems through disk capacity alarms.

Thanks @Penghui. In Kafka it was hard to find a health indicator that worked for everyone as when transactions came into play, some of the properties to indicate the health were stored on the server and it got complicated and confusing quickly. I need to dig into transactions a bit deeper to understand how that will play into this.