Our great sponsors
-
-
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
I found one related discussion online, which also seems to have someone running geth/prysm and getting missed/orphaned blocks specifically post-shapella. Their posted log entries show a similar pattern as yours, but even more extreme. Note the massive gap in time between payload work timeout at 08:13:59.743 and getHeader return at 08:14:08. That's more like 8 seconds late.
mev-boost sends out all its getHeader requests simultaneously, so if one relay said that it got it at 4897 ms into the slot, blocknative should have given the same error. Though blocknative runs their own relay implementation so maybe they haven't implemented a cutoff for getHeader. That's not great, they shouldn't be delivering a header if it's already too late to deliver a payload--and even then mev-boost shouldn't request the CL to sign a header if it's so late that all relays are going to reject it. That check should happen somewhere so that the proposer knows to try local block production, but it's not coded to do so.