Rate Limits - possibility to increase the threshold?

Dear all,

today, I switched my DNS provider - from now on, I would like to use desec.io

However, I found out the hard way that there are rate limits in place which makes updating a challenge.

My provider offers me a dynamic IPv6 prefix, and after getting a new prefix, I wish to update my IPv6 addresses for the servers in my network.

Is there a possibility to increase the number of possible updates per minute to maybe 20? As this will only happen whenever my prefix changes, the impact to the overall system will still be small.

Kind reagrds,

Hi,

you could look into bulk operations, which enable you to update multiple RRSets in a single operation.

That is against RIPE recommendations and you should switch your ISP if possible.
I know this isn’t always easy. This is just to say that we should not need to make changes to desec.io, just because some ISP is a fool.

On a more practical level, how about you randomise or stage the updates?

What do you use to for DynDNS?

You could my script on each host GitHub - jameskimmel/deSEC_DynDNS: DynDNS for deSEC.io · GitHub

and use crontab like

2-59/5 * * * * /home/YourUserName/deSEC_DynDNS.sh > /dev/null

3-59/5 * * * * /home/YourUserName/deSEC_DynDNS.sh > /dev/null

and so on

Hi Kornelius777,

Thanks for your message, and welcome to deSEC! :slight_smile:

The rate limits are there for a reason, as they are the most expensive operation in our API. But, that doesn’t make updates hard – see below.

The canonical way to do that is to use our dynamic update API, where the docs say:

In some cases, it is useful to only update the network prefix of existing A or AAAA records. To achieve this, append the prefix length to the IP address query parameter, such as ipv6=2a01:a:b:c::/64, myipv4=1.2.3.0/24, etc. The host part of the value is ignored.

Stay secure,
Peter

Whether dynamic prefixes, which complement the privacy extensions for the interface identifier, are a good idea or not is a controversial topic. The document you referred to is RIPE-690. It bases its persistent prefix recommendation on the technical difficulty of renumbering, which is necessary because in the absence of NAT, the hosts need to know their actual public IP addresses. There are other considerations besides this highly technical view, and calling ISPs who come to different conclusions “fools” is not helpful.

I am fine with giving the customer the option to choose.

But just having it dynamic, is indeed very stupid. Not all firewalls for example can even handle that.

I’m a bit confused, I use ddclient to update my IP address and while I don’t look up how often my address changes, but I know it doesn’t update my address when the address did not change. So it shouldn’t collide with any rate limits. I hope.

I thought this was the regular behaviour for those updaters. I used a different one which would perform a dns lookup and compare it with the current IP addresses. Result was the same, don’t update when the address hasn’t changed.

This topic doesn’t matter if you update just one hostname when your addresses change.

The problem arises in a situation where a dynamic IPv6 prefix changes and the user has lots of hosts with individual IPv6 addresses that all need to be adjusted when that happens. If you do this naively, for example with a ddclient for each host, then you’ll run into rate limits. DeSEC offers syntax to only update the prefix and keep the existing host part of an address, which allows you to change the prefix of many hosts in one request.

Ah, okay, I misunderstood the issue. I have multiple hostnames but all are being used on the same machine. So I guess, I won’t even run into this issue. Would be good if those rate limits wouldn’t be per account but per hostname.