I shall explain with the same example . Suppose a large organization wants to keep it 2 app DBs ( DB1 and DB2 ) in the same cluster . Over a period of time the organization realizes that the DB2 requires more compute power and storage . In order to add extra resources a budget needs to be allocated by the IT department handling DB2 .
Q1 . How can the sizing be done so that all new resources allocated be assigned to DB2 only , not DB1 ?
Q2. How can the new resources added be alloacted to just DB2 ?
Q3 . Is there a way we can identify the consumed resources by either DBs?
Thanks for sharing that. Some of our customers have addressed this, at least in part, by creating databases with different numbers of partitions. If you want a database to have access to less CPU resources, you can create it with fewer partitions, since currently we allocate one thread per partition. Since 7.3, we spread databases around the cluster when you rebalance.
Eric is referring to the partition count in the create database statement:
We have a future feature in the works to help out with this scenario. The ability to split up compute in a cluster into workspaces and attach different databases to different workspaces. We don’t have a precise ETA for this feature yet.