Wow, such amateur hour. I always thought that ntppool is run by competent people, and that each server has at least a GPS time source. They don't even seem to have a monitoring solution which can track what time each server is advertising.
I would switch to time.nist.gov if they had any servers in Europe. Any other good NTP EU source?
It's a free service, with volunteers providing the servers. There is monitoring, but as far as I'm aware, common ntpds don't relookup the configured servers while running, so you're going to be stuck to servers that tested good when your server started.
The downside of nist.gov servers is that they're sometimes very overloaded (and the one hosted in Washington state tends to have network path asymmetry, resulting in a large time offset reported).
> and the one hosted in Washington state tends to have network path asymmetry, resulting in a large time offset reported
The NTP algorithm compensates for delays in network traffic. It will result in a less accurate time, but it does not cause an offset in the resulting time.
> They don't even seem to have a monitoring solution which can track what time each server is advertising.
Yes they do. http://www.pool.ntp.org/scores/ In fact the page linked to even specifically mentions removing bad servers from the pool in real time.
> I would switch to time.nist.gov if they had any servers in Europe.
You talk about "amateur hour" but are seemingly unaware that unless you are running an NTP server for a significant number of users yourself, you should not be using time.nist.gov. Given that you're apparently currently using the pool, I'm guessing you're not.
Regarding time.nist.gov, it's still there in the list of Windows 10 NTP servers that a user can select, albeit, by default it will only be polled once per week.
Assuming something (that you care about) works a certain way without verifying the assumption is amateur hour.
Why would NIST have servers in Europe? Almost all of the big national standards institutes have publicly available time servers, NPL for UK, INRIM for Italy, etc.
Of course it tracks what time each server is serving (and other meta data). That's how the system knows that many servers didn't advertise the leap second when they should have (at which time it's too late for the NTP Pool to do anything useful about it).
Reportedly several national time sources (not NIST) managed to not announce the leap second, too.
I would switch to time.nist.gov if they had any servers in Europe. Any other good NTP EU source?