Currently, the advised way of deploying Readyset is by adding a read replica in the same hostgroup as a fallback and assign it a low weight. This still eventually redirects queries to read replicas, which is not desired.
Add a fallback hostgroup, where users can indicate a hostgroup which the scheduler will promote/copy the servers in case all readyset nodes are down.
Currently, the advised way of deploying Readyset is by adding a read replica in the same hostgroup as a fallback and assign it a low weight. This still eventually redirects queries to read replicas, which is not desired.
Add a fallback hostgroup, where users can indicate a hostgroup which the scheduler will promote/copy the servers in case all readyset nodes are down.