Description:
There's an interesting document here: https://dev.mysql.com/doc/refman/8.0/en/group-replication-throttling.html related to how GR auto-throttles writes in a way to ensure that it does not have problems. There's an additional page which mentions various metrics here: https://dev.mysql.com/doc/refman/8.0/en/group-replication-probes-and-statistics.html
To understand the throttling behaviour better and if it's being triggered it would be good to describe in more detail how the values specified are used, how they should be compared between different group members to allow the administrator to understand better what is limiting the throughput of the cluster.
Additionally perhaps some graphs could be shown to show a cluster which is throttled vs one which is not.
How to repeat:
Read the docs. You might like to know how to recognise if such throttling is taking place but not be sure quite what to do with the numbers specified.
Suggested fix:
Improve the documentation and provide some examples of metrics from a throttled system vs a non-throttled system.
If you get to the point where the system is self-throttling provide ideas on the possible options there are to improve the cluster throughput.
