Handling Millions of Records in Redis – Best Strategies?

Question

I store millions of user feed records in Redis under one key. What's the best practice for structuring and retrieving large datasets efficiently?