Comment 6 for bug 1881596

The use of when_any in the shared interface is causing the original bug.

    @reactive.when_any('endpoint.{endpoint_name}.broken',
                       'endpoint.{endpoint_name}.departed')
    def departed(self):
        flags = (
            self.expand_name('{endpoint_name}.connected'),
            self.expand_name('{endpoint_name}.available'),
        )
        for flag in flags:
            reactive.clear_flag(flag)

This update to the mysql-shared interface duplicates a change put into mysql-router which resolved scale in issues with mysql-innodb-cluster:

https://review.opendev.org/748717