Implement engine_getBlobsV1

[philknows]

Problem description

Reference: ethereum/execution-apis#559

Recently there have been calls to raise the blob target/max parameters for scaling DA. However, taking into consideration that some bandwidth/resource constrained validators are reportedly at their limits, we need to find ways to help alleviate this constraint before adding more demands for the most honest type of validator: home stakers building local blocks with 6 blobs and limited upload bandwidth.

Solution description

It's likely alongside Gossipsub 1.2 features such as IDONTWANT and APIs like engine_getBlobsV1 could help alleviate some bandwidth concerns so we can get better information about where blob limitations currently stand.

With #7077 already testing Gossipsub, we should prioritize supporting this as some execution clients have already implemented this feature for use.

Additional context

Teku analysis: https://hackmd.io/@dsjdqQDBRp66btjRzfxjbw/r1OuFG7AA
Lighthouse blog describing this optimization: https://blog.sigmaprime.io/peerdas-distributed-blob-building.html

[philknows]

Spec merged!

[philknows]

Will we be able to include metrics here @g11tech that capture how many blobs were retrieved from the mempool out of the ones you actually tried to retrieve over the wire? I think this might be useful to data collection, especially since we also produce a lot of local blocks due to our min-bid.

[nflaig]

hat capture how many blobs were retrieved from the mempool out of the ones you actually tried to retrieve over the wire?

I would expect this to have a higher impact on low bandwidth nodes and improve block import times there, potentially improving head vote accuracy / timeliness.