On Tue, Nov 04, 2025 at 02:50:37PM +0100, Grégoire Détrez wrote:
Hi,
Sorry, I think I made a mistake when computing the vkey (wrong key type) Here's the correct one (hopefully!):
witness.stagemole.eu+67f7aea0+BEqSG3yu9YrmcM3BHvQYTxwFj3uSWakQepafafpUqklv
$ echo "witness.stagemole.eu+67f7aea0+BEqSG3yu9YrmcM3BHvQYTxwFj3uSWakQepafafpUqklv" | sigsum-key from-vkey -v Key name "witness.stagemole.eu", key type: 0x04 (cosignature/v1) ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIEqSG3yu9YrmcM3BHvQYTxwFj3uSWakQepafafpUqklv sigsum key
Looks well-formed and with the right type!
On Mon, 3 Nov 2025 at 18:58, Rasmus Dahlberg rgdd@glasklarteknik.se wrote:
Hey,
Starting with the testing list sounds good (staging is not open yet).
https://testing.witness-network.org/log-list.1
We need to know how often this list is configured, e.g., hourly. Can you also update your witness about page with the info outlined here:
https://witness-network.org/participate/#specify-an-about-page-url
Sorry, I missed that. I'll fix it.
I'm intending to add
| Mullvad VPN AB | [testing/log-list.1][] | https://witness.stagemole.eu/about |
to the testing witness table here:
https://witness-network.org/witness-tables/
Just waiting for the about page to be ready -- so let me know when it is!
Regarding performance profiles: the idea is the witness operator picks a list they can support capacity wise. For example, if you have a witness that can only sign one checkpoint per second it wouldn't work to be on a list with a 10qps performance profile. But if you have a witness that handles 20qps, we (the maintainers) would expect 10 qps reserved for us.
So to answer your question, yes it defines what logs you get. We will basically only sign up as many logs as the performance profile allows!
(We don't have a performance profile for testing, but we will have that for both staging and production to maintain different lists reliably.)
Does that make sense and clear things up?
I think I understand the concept of the performance profile, but it was not clear what choices are available to me as a witness operator. But I see now that the staging list is named something like log-list XqpsYlogs. So I guess the pair (X,Y) is the performance profile and that there'll eventually be more than one?
Yep, correct. And I'd say a likely way this turns out is if we, e.g., have a 1qps list then the 10qps is probably having the 1qps list as a subset. So basically, as a witness operator you'd select the highest profile your able to accomodate, and we as maintainers will do our best to direct logs to the different lists without hogging all the capacity. So for example, we would definitely not put a 1qps log on the 1qps list, since then it's full. :-)
I think this will become a bit more clear once we actually have one (or maybe two) staging lists with different performance profiles. That said, happy to take MRs for the site with improvements if anyone has some free cycles!
https://github.com/transparency-dev/witness-network/tree/main/site
Otherwise we'll get onto improving the site more in a bit -- making a note that performance profiles weren't obvious when reading the site right now!
-Rasmus
Grégoire