What you need is an on-event hook..(called "pre-processor" in BE) This is the place where you can load necessary tuples from the store. Also this needs to be integrated with the rule session api with a "Store" interface.
Incidentally I have a WIP branch that addresses the store interface and some more (like an interface to externalise join tables themselves and an in-memory implementation of those) Perhaps you can use that as the starting point for implementing your Solr/Redis store?

Take a look at the 'store' branch to start with.

@balamg Sure, will take a look.

After reviewing the WIP store branch, it would be best to merge the two branches first. The merge would include a part of:

• redis tuple store with update support
• TupleType based secondary index in redis
• multi store support(per tupletype store)
• pre-processor(or startup) time tuple loading
• tuple update propagation to clustered engines.

Thus closing the PR.