Partition cell_outputs table

[ShiningRay]

The cell_outputs tables has more than 60,000,000 records.
The RDBMS usually has performance drop significantly after a table exceeds 20,000,000 records.
This is because the algorithm of B-Tree index, which has limit according to the layers in B-Tree.
So we can utilize the "partition" mechenism of RDBMS, to split big table into small partitions while not affecting the higher codes.

[Keith-CY]

By 11/09

[Keith-CY]

Blocked by migration of k8s

[Keith-CY]

Waiting for support of @mdzh521

[ShiningRay]

Suggested partition column:

• cell status: active or dead (maybe pending when ckb_transactions and pool_transaction_entries are merged)
• range(epoch) ( missing epoch column, add a epoch column )
• range(block_number)
• hash(transaction_id)