feat: larger batches for EigenDA

[epociask]

Changes

• Support max batch size config for eigenda integrations with 2 MB upper limit

Dependency PRs

• Epociask feat larger batches nitro-testnode#6

Testing
Used testnode's spam feature to generate arbitrary L2 txs using 1096 bytes of calldata across 3 address ran across 3 processes:
i.e, sequencer --l2_tx-> validator --l2_tx-> l2_owner --> EOA

Using a testnode policy where batches would only submit every 15 minutes, the txs were sent serially across 20 threads per process. In the best case this would equate to roughly 60 txs / second (i.e, 60 * 1,096b = 65.760 kb/sec) or 5.918 mb/batch. Since these scripts use asynchronously concurrency bound by js and serially execute the real throughput expressed during this experiment is likely a bit less. Anyway, was able to get batches through around sizes of 780,328 bytes . The stateless block validator struggled very heavily to retain synchronization given it'd only sequentially execute 10-20 messages in a batch with the number of incoming messages drastically increasing:

INFO [08-07|04:14:32.927] validated execution                      messageCount=6224 globalstate="{BlockHash:0x85499c64e0b9ab24d5d7c094108c5fc30c61389b222db771b50cee5ffccff876 SendRoot:0x0000000000000000000000000000000000000000000000000000000000000000 Batch:2 PosInBatch:3127}" WasmRoots=[0xe82561d483d1b87cf09343ec6a1b9bf31b98ff4ab2b26b444f7d4eadb353a314]
INFO [08-07|04:14:33.927] validated execution                      messageCount=6250 globalstate="{BlockHash:0x1639290819a088772592f0fcbd96e0a82dbf03071f8e359d89f12a1d0f50aae4 SendRoot:0x0000000000000000000000000000000000000000000000000000000000000000 Batch:3 PosInBatch:0}" WasmRoots=[0xe82561d483d1b87cf09343ec6a1b9bf31b98ff4ab2b26b444f7d4eadb353a314]

This delta if left un-remediated would likely cause further delay to child --> parent chain withdrawal times.

TODO(s)
The following items should be completed to further verify that our fork can support batches of this size:
[] kzg-bn254 proof generation tests using 2MB blobs
[] experiment using 2MB max batch size

[bxue-l2]

when needed, we can put more srs up to 2^28

[hopeyen]

my first review! everything looks good, just left some dumb questions:)