On Wed, Nov 05, 2025 at 08:27:39AM +0100, Grégoire Détrez wrote:
On Tue, 4 Nov 2025 at 15:21, Rasmus Dahlberg rgdd@glasklarteknik.se wrote:
$ 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!
🎉 🎉 🎉
Starting with the testing list sounds good (staging is not open yet).
I was wondering about this: does it mean we'll need to pull both the test list and the (selected) staging list in the future?
Good question -- the test list is intentionally a bit YOLO so I'd say it's ok that you stop configuring it and instead switch to staging later. If anyone has a different take I'd be curious to hear it!
I'm intending to add
| Mullvad VPN AB | [testing/log-list.1][] | <https://witness.stagemole.eu/about%3E |
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!
Ready now ☑️
Just queued you up in the testing table -- you can see the status here:
https://github.com/transparency-dev/witness-network/pull/18
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. :-)
Thanks, that does make it clearer! (Sounds like a fun allocation problem 🤓)
Fantastic, thanks for asking the questions!
-Rasmus
Grégoire
Hi Grégoire,
We just started populating the witness staging table. I had a local note to revisit this email thread once this happened.
Are you still interested to transition your witness from testing to staging? If yes, I can file a PR that gets this transition done.
Concretly what I'd suggest is:
* Keep the testing list configured (this is similar to how, e.g., Google's staging witness is configured: both test and staging). * Also configure the staging log list <-- this needs to be done before I can file my PR, so let me know if/when it's ready.
Thanks for contributing :-).
-Rasmus
On Wed, Nov 05, 2025 at 07:17:49PM +0100, Rasmus Dahlberg wrote:
On Wed, Nov 05, 2025 at 08:27:39AM +0100, Grégoire Détrez wrote:
On Tue, 4 Nov 2025 at 15:21, Rasmus Dahlberg rgdd@glasklarteknik.se wrote:
$ 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!
🎉 🎉 🎉
Starting with the testing list sounds good (staging is not open yet).
I was wondering about this: does it mean we'll need to pull both the test list and the (selected) staging list in the future?
Good question -- the test list is intentionally a bit YOLO so I'd say it's ok that you stop configuring it and instead switch to staging later. If anyone has a different take I'd be curious to hear it!
I'm intending to add
| Mullvad VPN AB | [testing/log-list.1][] | <https://witness.stagemole.eu/about%3E |
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!
Ready now ☑️
Just queued you up in the testing table -- you can see the status here:
https://github.com/transparency-dev/witness-network/pull/18
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. :-)
Thanks, that does make it clearer! (Sounds like a fun allocation problem 🤓)
Fantastic, thanks for asking the questions!
-Rasmus
Grégoire
participate@lists.witness-network.org
-
Rasmus Dahlberg