When handling factories, the realtime sync currently queries the database on every block to determine which logs or traces are matched by a filter that contains a factory.

We could skip this database query if instead, on startup, we queried all child addresses, stored them in memory, and updated when a new child is registered.