Feature request: Expose NTP metrics to monitoring infrastructure

[scoopex]

As discussed in #1852 it would be useful to expose the details of the ntp time synchronization state (e.g. offset, jitter, delay, count of "good" servers) to the monitoring infrastructure.
In my understanding SONiC itself currently only exposes metrics by SNMP.

As @saiarcot895 pointed out, it might be a good idea to expose the information to STATE_DB periodically (i.e. 5 minutes) where it can be consumed also from a prometheus exporter.
Probably this script might be a source where you can get a first draft of this functionality.