The stratum is only one piece of information NTP will use in determining it's source. NTP is choosing to use the local time source because it has deemed it 'better' (or more trustworthy) than the other sources you have configured. ISE recognizes it as stratum 4 server, but does not choose it as time source. Just to make sure it is not an NTP server issue, I configured one of our Cisco routers as NTP server, which itself uses 146.140.66.230.

Additionally, as far as I know, none of both sersers is a Microsoft NTP. As I do not have access to the root shell, I cannot use ntpq as indicated by this tech note. The offset however is roughly correct, slightly more than 2 minutes. Interestingly the offset differs quite significant looking at both time servers. Warning: Output results may conflict during periods of changing synchronization.Д«oth of them are reachable (the first one even in the same subnet), as can be also seen at the delay / jitter values. * Current time source, + Candidate, x False ticker Remote refid st t when poll reach delay offset jitter I configured the same NTP server on my Cisco ISE, but even after multiple restarts it does not use it as NTP source. All switches, firewalls and other components use it as timezone and it works. To have synchonized times, we use an NTP server in the same subnet as the ISE to sync all the components (stratum 3, based on Bluecat appliances). We have a Cisco ISE 2.0 in our lab doing some 802.1X testing.

Cisco ise 2.4 0 patch 5