Allow retrying sampling of a data column on a peer after certain amount of time

[jimmygchen]

Description

#6084 contains a change for Lighthouse to not retry fetching the same sample from the peer if it's not able to return to us, and we just continue fetching from the next available peer. This is fine in most cases, however if we have very limited number of peers, it may be useful to retry again, as not having a sample is could be a temporary state.

See comment here: #6084 (comment)