With electra, the getHeader response now contains all the execution requests, making the worst case size of the response ~5MB, which can add significant latency on the hot path. There also has been concerns (eg. #108) about leaking block information via the logs bloom.
Would be great to assess whether we can strip down the response to the minimal subset needed for the proposer to properly validate the header and sign, thus reducing load / latency on the hot path
With electra, the
getHeaderresponse now contains all the execution requests, making the worst case size of the response ~5MB, which can add significant latency on the hot path. There also has been concerns (eg. #108) about leaking block information via the logs bloom.Would be great to assess whether we can strip down the response to the minimal subset needed for the proposer to properly validate the header and sign, thus reducing load / latency on the hot path