Using local clock mode is not recommended
This cookbook currently adds the local clock as a high-stratum time source on non-virtualized, non-Windows hosts.
In the event that NTP servers are unreachable, this will cause every host to sync to its own clock, and machines will drift apart. This also causes any monitoring to think that the host is synchronized when it's not.
In the past, it was recommended practice to configure one server with a local clock source to provide redundancy if external servers are unreachable. This is no longer recommended practice, and NTP orphan mode should be used instead.
@russss I'd be open to a PR that would align us with the best practices if you have the time
I've hit situations as described where a host cannot reach an upstream ntp server and falls back to the local lock and unless you really dig deep into ntpq query information it appears synced.