[RFC] Using project and service specific DNS end points to support flexible deployments.
DJ Delorie
dj@redhat.com
Tue Nov 10 01:52:53 GMT 2026
(sorry, the send button is next to the expand button)
On Sat, Nov 7, 2026 at 10:51 AM Carlos O'Donell <carlos@redhat.com> wrote:
> > All I mean is that if we do "project.host/service" we have one
> > ingress/dns path, and we can use internal reverse proxies to manage
> > service load and per-service dns without the users even knowing (aside
> > from removing an ingress point).
>
> This...
>
> > With service.project.host/ you need an ingress concentrator for each
> > service, and service dns changes are public-facing and thus affect
> > users and need to be scheduled much further in advance. Plus you need
> > to manage far more public DNS entries - number of services times
> > number of ingress points.
> ... and this are the same.
Only when they're on the same machine.
Imagine the worst (best?) case where each endpoint has a dozen
independent hosts serving its user-facing interface, and a dozen
independent hosts running its backend.
That means you need a dozen ingress hosts for bugs.glibc.*, a dozen
for git.glibc.*, a dozen for mail-archives.glibc.*, etc. Each needs
management, certs, dns, etc.
Or... glibc.* has two dozen ingress hosts that are identical.
> The use of service.project.host does not *require* a separate ingress
> concentrator per service, but it permits it.
Right. It's in the "permitted" configuration that maintenance of the
servers becomes more difficult, because you have so many different
configurations.
> In path-style the route is keyed at the reverse proxy by URL path.
>
> In subdomain-style the route is keyed at the reverse proxy by HTTP Host header.
>
> You can handle both in the same reverse proxy AFAIK.
Again, only when you're on one machine. I'm planning for running on
lots of machines, world-wide. If we plan for one machine, there's no
point breaking them apart at all.
> A wildcard record *.glibc.sourceware.org is one DNS entry.
This only works if you're sharing ingress servers, and if you are, you
might as well use an url-based route. Having a subdomain-based route
is only a benefit if you have different DNS entries for each service.
> Origin isolation for anything browser based puts cookies in different
> security domains. If you don't do this then a bad services that has
> an XSS can read another's cookies and use those credentials to do
> something with a service that matters *more* than one of the side
> services.
If this is your concern, you shouldn't be using wildcard certificates.
You'd need to be able to revoke a compromised certificate if a bad
actor gets into one of the services.
With an url-based route, the certificate is handled by the ingress
server, and a bad actor in the *service* doesn't have access to it.
Of course, you can do this with any ingress/backend pairing.
More information about the Libc-alpha
mailing list