[Feature] Batched writes during subgraph indexing

[azf20]

Description

Currently Graph Node writes entities to the store on every block. This leads to a large number of database interactions during historical syncing. This can be a significant performance bottleneck.

Graph Node could instead store entity changes in memory, writing those changes to the database every X blocks, or when the diff to be written is above a certain size.

This should combine with #4477 to reduce the number and size of database updates during syncing.

Are you aware of any blockers that must be resolved before implementing this feature? If so, which? Link to any relevant GitHub issues.

• Impact on proof of indexing (ensuring that POI is not impacted by this change)

Some information to help us out

• Tick this box if you plan on implementing this feature yourself.
• I have searched the issue tracker to make sure this issue is not a duplicate.

[neysofu]

Closed by #4606.