Summary

The Hivemind cluster initially sat and operated alongside our central Collective shared infrastructure cluster to intentionally "run ahead" of the hive version included in ACM's GA-ed versions to pick up new clusterpool-centric changes.  This environment remained relevant to run ahead and encounter new features and API changes in Hive before they reached the ACM codebase.  We caught a number of breaking changes via this method.  

Now that hive is exiting rapid development, we haven't seen as much benefit from this run-ahead hive version.  Also, we're working and integrating more tightly with the hive team, to encounter changes earlier.  We've also seen increasing overhead to maintain two redundant clusters, which are both quite old.  

The time has come to migrate CI load off of Hivemind to Collective, verify migration and stability, and tear down Hivemind.  

Goals

1. Migrate CI pools to Collective and off of Hivemind - including reconfiguring CI to pull resources from Collective
2. Verify migration via a soak test
3. Deprovision Hivemind