I would like to have at least some level of support for Group Replication set-ups in Orchestrator. The bare minimum support would include:

• At least the (simpler) single primary mode, since it is somewhat more similar to the traditional single primary master in async/semi-sync replication.
• The API returns all group members as part of the same cluster in calls to /api/cluster/<cluster_alias>
• The web UI shows the group members in the same view, instead of what happens today, in which each group member looks like a separate cluster.
• Orchestrator does not consider GTIDs from the group as errant anymore.
• Orchestrator prevents the user from performing incorrect topology for group members (e.g. set a primary to replicate from a secondary; or move a secondary to replicate from somewhere else, since secondaries actually replicate from the group's primary).