Op 27-08-2026 om 13:55 schreef Daan Hoogland:
On Thu, Aug 27, 2026 at 10:12 AM Wido den Hollander
<[email protected]> wrote:
Op 27-08-2026 om 09:52 schreef Daan Hoogland:
... a lot withbad spelling
Thank you for the feedback! So I want to keep the data as anonymous as
possible, only counts of types. No hostnames, no references.
agreed
...
Also a collector app would be needed (I propose Claude and I write it
in rust ;) which creates a data pool.
It's the already, a Python application which runs as WSGI and stores all
the incoming reports as JSON files. From there we can process them.
Great, so the processor would be a webclient loading the json.
Python receives it and stores JSON files. Design for now. Afterwards we
can use anything to process these files and make something out of it.
Initially not a lot of people will have this turned on, so the data will
trickle in.
As for access to highly anonymised, denormalised, aggregated data. I
see not much harm in publishing it on a publicly accessible dashboard.
Me neither, it would be super transparant. We would then need to
motivate people to turn this setting on.
We would also need to request the ASF/Infra to create a subdomain for us
under the cloudstack.org domain name.
Do you mean xxx.cloudstack.apache.org, because I don't think apache
has control over cloudstack.org, do they?
They do. I already created a ticket [0] and INFRA created a record very
quickly, did not expect that to happen.
call-home.cloudstack.org is a CNAME to CDN77 now, they also sponsor us
for download.cloudstack.org
There is no backend yet. The code needs to be checked and we need
consensus if we want to have this in the management server to start with.
More input is welcome!
Wido
[0]: https://issues.apache.org/jira/browse/INFRA-28328
Wido
[0]: https://github.com/apache/cloudstack/pull/13985
thanks for reviving this, Wido!
