Hello Elljay,

These are business, the ones impacted have between 20 and 50 devices that will share a single IP (NAT). I have monitoring running on each device. The monitoring makes a single request from each device every 4 hours. If the request fails on a device 3 times in a row, I'm alerted. So some machines are giving me errors, some are fine. (because sometimes a request works).

I am able to reproduce this from my home though. It's also a Telenet internet (not business) line, but obviously, not many devices will be making requests over here.

So in general, I'm making a single packet request every 4 hours from each device. However, the TCPdump and when I manually test is using the w32tm command from windows. This requests an update each second. (again, that's only to test and that's just for a few seconds).
This is the result this morning from my own device at home (so no other devices requests NTP over here)

Single problematic server from the pool:
w32tm /stripchart /computer:45.87.76.3 /dataonly
Tracking 45.87.76.3 [45.87.76.3:123].
The current time is 14/09/2019 09:55:06.
09:55:06, error: 0x800705B4
09:55:09, error: 0x800705B4
09:55:12, error: 0x800705B4

be pool:
w32tm /stripchart /computer:be.pool.ntp.org /dataonly
Tracking be.pool.ntp.org [82.64.45.50:123].
The current time is 14/09/2019 09:56:03.
09:56:03, +01.7553524s
09:56:05, +01.7559566s
09:56:07, error: 0x800705B4
09:56:10, +01.7558103s
09:56:12, +01.7560005s
09:56:14, error: 0x800705B4

Belnet:
w32tm /stripchart /computer:ntp.belnet.be /dataonly
Tracking ntp.belnet.be [[2001:6a8:3c80::14]:123].
The current time is 14/09/2019 09:57:03.
09:57:03, +01.7556700s
09:57:05, +01.7562380s
09:57:07, +01.7577016s
09:57:09, +01.7554397s
09:57:11, +01.7553191s
09:57:13, +01.7572620s
09:57:15, +01.7565029s

Trace to first server:
tracert 45.87.76.3

Tracing route to ntp.devrandom.be [45.87.76.3]
over a maximum of 30 hops:

1 2 ms 2 ms * 192.168.126.1
2 15 ms 14 ms 12 ms dd576a001.access.telenet.be [213.118.160.1]
3 13 ms 13 ms 13 ms dD5E0C969.access.telenet.be [213.224.201.105]
4 * * * Request timed out.
5 21 ms 21 ms 33 ms 213.46.174.101
6 23 ms 38 ms 23 ms ae-8.r04.londen12.uk.bb.gin.ntt.net [129.250.66.101]
7 22 ms 20 ms 23 ms ae-3.r25.londen12.uk.bb.gin.ntt.net [129.250.4.128]
8 28 ms 27 ms 24 ms ae-2.r24.amstnl02.nl.bb.gin.ntt.net [129.250.5.2]
9 27 ms 26 ms 26 ms ae-7.r03.amstnl02.nl.bb.gin.ntt.net [129.250.2.103]
10 76 ms 139 ms 32 ms ntt2-eqx-am7.r2.nzs.nl.as43350.com [81.20.72.105]
11 30 ms 27 ms 27 ms nforce.router01.stb.devrandom.be [185.107.81.97]
12 31 ms 27 ms 27 ms ntp.devrandom.be [45.87.76.3]

Trace complete.

Note that I have no ICMP loss when performing echo requests.

But this is where is gets weird. I have IPv6 at home. So have a look at this:

>ping ntp.devrandom.be -4

Pinging ntp.devrandom.be [2a0e:f780::3] with 32 bytes of data:
Reply from 2a0e:f780::3: time=25ms

Ping statistics for 2a0e:f780::3:
Packets: Sent = 1, Received = 1, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 25ms, Maximum = 25ms, Average = 25ms
Control-C

>ping ntp.devrandom.be -4

Pinging ntp.devrandom.be [45.87.76.3] with 32 bytes of data:
Reply from 45.87.76.3: bytes=32 time=27ms TTL=56

Ping statistics for 45.87.76.3:
Packets: Sent = 1, Received = 1, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 27ms, Maximum = 27ms, Average = 27ms
Control-C

>w32tm /stripchart /computer:2a0e:f780::3 /dataonly /samples:10
Tracking 2a0e:f780::3 [[2a0e:f780::3]:123].
Collecting 10 samples.
The current time is 14/09/2019 10:11:25.
10:11:25, +01.7616244s
10:11:27, +01.7565392s
10:11:29, +01.7591802s
10:11:32, +01.7574163s
10:11:34, +01.7547778s
10:11:36, +01.7562617s
10:11:38, +01.7570559s
10:11:40, +01.7563072s
10:11:42, +01.7579061s
10:11:44, +01.7565173s

>w32tm /stripchart /computer:45.87.76.3 /dataonly /samples:10
Tracking 45.87.76.3 [45.87.76.3:123].
Collecting 10 samples.
The current time is 14/09/2019 10:11:48.
10:11:48, error: 0x800705B4
10:11:51, +01.7585152s
10:11:53, error: 0x800705B4
10:11:56, error: 0x800705B4
10:11:59, error: 0x800705B4
10:12:02, error: 0x800705B4
10:12:05, error: 0x800705B4
10:12:08, error: 0x800705B4
10:12:11, error: 0x800705B4
10:12:14, error: 0x800705B4

The pool server is working just great on IPV6, horrible on IPV4.