So much for the Fritzbox, 3 days stable and it started restarting every few minutes.
There is something seriously wrong with 07xx and 08xx latest firmwares.
I flashed back to old stable (non-labor) 07.62 firmware to see if it starts again.
The systemlogs gave me DSL adjustment messages BUT DSL is disabled in the Fritzbox. 
Itâs connected via the WAN connector to the Zyxel that does the DSL-part.
Fingers crossing again that it keeps stableâŠelse I realy need to build me a router.
In the current monitoring system the monitoring, scoring and selection of âactive monitorsâ are loosely coupled. For operating the system this works great, and I believe it keeps each of the pieces simpler, but it does miss the feature of graphing based just on active monitors.
Iâve been tempted to add a flag in the monitoring data to indicate if the monitor is âtestingâ or âactiveâ, but it doesnât feel quite correct because the âtestingâ monitors can influence scoring; and the selection and scoring happens (very shortly) after the monitoring data is recorded.
Anyway, I take this thread as a feature request for better graphs. 
Help is welcome. Doing the JavaScript / graphics work will likely take forever for me so I havenât touched that code for many years. Thereâs a reasonable API for fetching monitoring data now, including old data and it can easily be extended.
@ask, it would be good if we can set a monitor in testing ourselves, in case of problems like I have at the moment with my modem(s).
As I am working hard to resolve it, but itâs not easy when it happens only after 3 days.
And I have downgraded the firmware and prepared a replacement router based on IPFire on an Advantage ARK1124 that I have.
However, this means I have to reconfigure my entire network, that is a lot of work, as such that is my last resort.
The problem at the moment isnât speed but rebooting Fritzboxes and with wrong messages in the logs. AVM is really making a mess of their latest firmwares.
I will be stable and solid againâŠthat is a promise.
However, I need the monitor to run as without it the problem takes a lot longer to happen.
I was planning to install IPv6 again today, but the reboots started and wouldnât stop. I had to remove the Fritzbox from the modem in order to stop it, that bad.
Well itâs running for 2 days now without slowdowns, reboots or any problem.
My assumption that AVM made a mess of new firmwares after 07.8x is correct, just look here:
https://forums.thinkbroadband.com/zen/4765100-fritzos-800.html
Out of the many pages where people complain about new FritzOS where routers reboot, disconnect, become slow etc.
It seems dat any FritzOS before (my best guess) 07.80 is ok, but after itâs unstable, regardless if itâs a release or not.
I complained to AVM about my 7590 6 months ago, they even replaced the box out of warrantyâŠwhen I told them the ânewâ box had the same problemsâŠthe answer was: Your box is out of warrenty!
I did point out to them that the 7530AX had the same problems and was brand new.
Then they stopped answering my messages. As it looks now, nobody at AVM steps up and says: We made a mistake! And revoke all new firmwares and offer the latest GOOD firmware.
But they donât. And they do not even offer old firmware recoveryâŠso if you look for it:
https://avm.ottertel.de/fritzbox/
and here:
Beware, a backup of a newer firmware can NOT be restored to older firmware.
I flashed back:
7590 - 07.59
7530AX - 07.57
5960Pro - 07.62
And all of them seem stable, no slowdowns, reboots or other problems.
Fingers crossed 
3 dayâs up and no issues.
I have put the Fritzbox 5690Pro under real stress now, downloading (multithreading) ISO-CDâs so it does max out the line as much as possible.
As 123 has priority over it, it should not affect my monitor.
Even the GUI of the Fritzbox stays fast, it never did that with the newer firmwares.
So much for AVMâs newer, better and improved firmwareâŠdonât know what those developers where sniffing when they released it, as itâs the source of the problems I was having.
Over a few hours when itâs finished downloading Linux distroâsâŠloads of them, I will enable IPv6 again.
Finally!!!
Update: IPv6 has been re-activated.
Please let me know if my monitor misbehaves, either IPv4 or IPv6.
It should behave normally for most.
Modem is still stable and no issuesâŠ
Update:
Unless you control the ISP router your router is connecting to, your prioritization of NTP affects only the monitorâs outbound queries. The server responses can still be queued behind the download traffic by the ISPâs customer-facing router.
And it took 2 daysâŠthe router rebooted.
IPv6 has been removed again.
For some reason IPv6 seems to overload my router.
It ran fine without IPv6.
Another few weeks to run without IPv6 and see how it goes.
Also the IPv6 monitor is turned off.
Hello,
I am interested in learning thatïŒwhether the quality monitor for the servers is sourced from the official NTP ïŒor from volunteer servers.If it is the latter,would it be possible for me to contribute my server as a monitoring server?
My server is only equipped with a 3Mbps bandwidth,which makes it utterly insufficient to withstand the NTP traffic surge in the Chinese region.However,fortunately,my server boasts a relatively good network connection,performing well when accessing most areas within China.Therefore,I am wondering if I could utilize my server as a monitor,which would address the issue of not being able to reach Chinese servers in the previous discussion.Any response would be greatly appreciated.
Best regards.
Yes, itâs possible.
Yes, another(?) monitor in China would definitely help avoid Chinese servers being kicked out of the pool due to issues isolated to international clients. 3 Mbps is plenty more than used by a pool monitor (really at least two monitors if you have both IPv4 and IPv6), as each active monitor sends no more than a few tiny packets per second currently.
However, be aware the monitoring is designed to choose the 5 highest-scoring monitors to be active monitors for each pool server. The median score from those 5 active monitors is the score used to decide if a server is included in the pool, so it wonât completely solve the problem. That would take ensuring there are at least 3 well-connected monitors in China to ensure the median (3rd place of 5) score reflects the view from within China, assuming the other two Chinese monitors are first and second among the five active monitors.
You need a Linux or FreeBSD system to run a monitor. It is also important that the server is not routinely heavily CPU loaded. To volunteer your server as a monitor, email support@ntppool.org. If you donât get a response in a day or two, I can refer you to other addresses. Iâm at davehart@gmail.com.
Remind me which router youâre using now? Iâd expect a FritzBox to handle IPv6 well as itâs so widely deployed in Germany and Western Europe in general.
Iâve seen consumer-grade routers in the US that handle IPv4 much better than IPv6 as weâve been relatively late to IPv6 except for cellphones thanks to our oversupply of IPv4 addresses going back to the days of, for example, Apple being assigned 17.0.0.0/8 with something like 1/220th of all IPv4 unicast addresses in that block. I think Apple still has all of that space, though I highly doubt theyâre using 16.7 million IP addresses on the internet.
With IPv4 addresses being leased by some for $10/month and probably sold for quite a bit more, Apple may have a little cushion of cash hiding in unused IP addresses, and would be financially wise to keep them as they keep getting more valuable over time.
When I started at Microsoft in 1991 every internal PC that was using IP was assigned a global IPv4 address despite only a handful of them being connected to the internet. Only a minority of PCs at Microsoft were using IP then, they mostly used XNS in those days as the transport protocol internally on one big happy L2 LAN with 100 Mbps redundant fiber links between buildings and 100 Mbps cat5 to servers, and 10 Mbps AUI connections to office and lab PCs, with a few dozen to a few hundred sharing one thicknet coax link the AUI connections were backhauled with. By the time I left in 1998 they had upgraded to switched 100 Mbps client connections and gigabit ethernet for servers but were still using global IP addresses thanks to their early adoption of the Internet getting them a large allocation gratis.