I take a big hit in queries

Yes, it is the combined traffic both for IPv4 and IPv6. I set the maximum speed settings available (3 Gbit).

Here it is the page for all the servers that I am maintaining: pool.ntp.org: NTPman's pool servers . The last one is testing only speed, that is a test monitor in the beta pool.

I told you it’s underserverd and being hit massive.

We should have the same number of servers as e.g. NL, but we don’t.

My numbers still going up:

be 	8.398 ‱ 	0.205 ‱ (40.88x)

Not as bad as before the extra help….thanks guy’s!:+1:

I find it hardly surprising that a server that used to serve only one zone (ch) gets ~double the traffic when another zone (be) gets added. If anything, it tells us that the supply/demand for NTP pool services is similar in Switzerland and Belgium and I haven’t heard of the Swiss complaining that their servers are being hit hard.

Focusing on the NTP queries again. Bas, do you think you are getting too many NTP queries at the moment? If yes, what would be an acceptable NTP query rate per second? If no, there’s no problem, right?

1 Like

There are a number of factors that are in play when determining if a certain traffic volume is problematic. We all are aware of them (BW, line type and line symmetry, hardware used etc).

At the moment there isn’t a problem. As others jumped in to help BE.

But a few weeks ago when I ran both IPv4 and IPv6, I got massive requests.

So I dropped IPv6, changed router etc. It’s just a few days that I enabled service again.

I can not say if it happens again. But so far the req/sec seem ok.

When my home-network slows again, I have to measure the req/sec and see how much it is.

It’s like a car, a mcheanic can’t fix the problem when the problem doesn’t show :grin:

When I started this topic it was slowing down my network.

2 Likes

How did you guy’s got their servers added to Belgium?

As I have 1 big server, doing IPv4+IPv6 that has 100mbit, 3CPU’s and 12GB ram, set at 12 and 10mbit traffic.

and

or short: ntp3.heppen.be

This may be added to Belgium. As this server has plenty speed and bandwidth to help Belgium.

How do you got it added?

I just sent a request to: server-owner-help@ntppool.org.

I now handle 5% of IPv6 DNS requests for Belgium.
and contributing also 5% for bandwidth for Belgium. Relative bw setting 1 Gbit.

1 Like

Thanks, request send :+1:

I told you, when entering our pool, the numbers will be massive.

That is my problem before. I got hammered.

But my German VPS server has not much to do, it can also help.

For now the numbers are set 10/12mbit, could be higher if it’s not overloaded.

It has an 100mbit connection.

hi Bas,

Was your request handled?

Nope, not yet. Give them a little bit of time.

Anyway, my number dropped further to : be 5.706 ‱ 0.205 ‱ (27.77x)

Edit: Email send again, but this time my forum/login email, that I normally not use.

Update: Yes my German server is added to BE for IPv4 and 6 :+1:

See when it was added:

 chronyc -n clients |sort -rn -k 3,3 |head

It will sort the output by the highest number in the third column.
But it works only if you haven’t disabled the client log.

Biggest is this one:

That is Telenet ISP in Belgium.

I have added a webpage to my home-server-strat-1 so I do not have to give much details all the time.

Updated every 5 minutes:

http://websdr.heppen.be:90/

Update for a weekly statistics as the servers were added to the be country zone at Wednesday evening:

One server pool.ntp.org: Statistics for 2a00:7580:60:211::46 has the following:

be 1311.59 ‱ 1609.31 ‱ (0.81x)

The other one pool.ntp.org: Statistics for 2a00:7580:60:211::52 :

be 1311.45 ‱ 1609.31 ‱ (0.81x)

What is the meaning of those two set of values accomulated expressed in % concerning the distribution of the time servers for the be country zone?

I was wondering that too, but so far no idea.

I only know if they are high you get lots of requests.

The second permyriad tells you that

  • for both your servers combined, the share of the overall sum of netspeeds was more than 32% at that time,
  • the overall sum of all netspeeds was about 18.65 Gbit at that time (assuming yours were both at 3Gbit),
  • and that based on the current number of servers in the zone (21, which might have been different at the time you took the screenshot), the average netspeed for the remaining 19 servers was ~666 Mbit.

The first permyriad tells you that aggregated over about the last three days, your servers combined were present in more than 26% of DNS responses to requests from the BE zone.

The ratio between the two can tell you that the IPv6 zone for BE is very likely rather well-served now.

2 Likes

Nope, not necessarily true. It depends on the traffic demand in the respective zone. For IPv4 in BE, it would indeed be quite a bit. But for IPv6 it is only a fraction of what one would see for IPv4.

Overall, I think it depends on the number of servers vs. the demand in the zone, i.e., number of clients and how they behave.

E.g., consider the following IPv6 server:

Screenshot from 2025-10-20 19-47-19

It’s share is even higher than the examples above, but it gets a meager average bitrate of typically slightly less than 15 kbit/s (typically less than 20 packets per second).

Sorry still no clue what it means.

If you were more specific as to what aspect you don’t understand, it would be somewhat easier to address that.

I do not understand the numbers, it makes no sense….high/low…requests? WTF do they mean?

There is no reference to x.xx or x.xx 0/000…it tells us nothing.

Maybe you have an idea, but I’m lost as there is no meaning in what is what.

In short….it makes no sense. Demand of what? What is the demand? What is the baseline?

What is a normal number? I mean…sorry? I’m utterly clueless.