Leader election times seem too long with max current term

I use a fault injection tool to modify the term field of AppendEntriesRequest to max uint64.
When the current term is 18446744073709551615 (max uint64) , restart current leader node to start a new election.
It takes a long time to select a new leader(about 1min).
The testing system is rqlite.
Some crucial logs can be find in https://github.com/rqlite/rqlite/issues/1712#issuecomment-1975546515

Start by tracing how AppendEntriesRequest handles a max-uint64 term and how that value affects leader election after the current leader restarts. Review the linked rqlite logs and relevant election behavior; done means a new leader is selected without the reported one-minute delay and regression coverage exists for the maximum term.

Written by the indexing model from the issue text.