discovery.etcd.io down "503 Service Temporarily Unavailable"

[yogeek]

Bug report criteria

• This bug report is not security related, security issues should be disclosed privately via security@etcd.io.
• This is not a support request or question, support requests or questions should be raised in the etcd discussion forums.
• You have read the etcd bug reporting guidelines.
• Existing open issues along with etcd frequently asked questions have been checked and this is not a duplicate.

What happened?

We rely heavily on "discovery.etcd.io" to manage our ETCD clusters with terraform and the site seems down currently so we cannot deploy new clusters anymore.

What did you expect to happen?

discovery.etcd.io available at any time

How can we reproduce it (as minimally and precisely as possible)?

go to discovery.etcd.io : "503 Service Temporarily Unavailable"

Anything else we need to know?

I find a few things related :

• https://etcd.io/blog/2025/announcing-etcd-3.6/#v3discovery
• What's the future of discovery.etcd.io discovery.etcd.io#82
  but it does not give a clear status about the future of the site.

Etcd version (please run commands below)

3.5.24

Etcd configuration (command line flags or environment variables)

No response

Etcd debug information (please run commands below, feel free to obfuscate the IP address or FQDN in the output)

Relevant log output

[al-ti]

We are affected by this as well.

[ahrtr]

@yogeek @al-ti are you depending on discovery.etcd.io to bootstrap your production etcd clusters?

As we already announced in https://etcd.io/blog/2025/announcing-etcd-3.6/#v3discovery, discovery.etcd.io has been deprecated. It's recommended to use v3discovery. You can deploy an etcd cluster to bootstrap other etcd clusters.

[yogeek]

Yes we do, because it was really useful to have a public endpoint to create a cluster from scratch and to avoid the chicken-egg issue you get when you need an etcd cluster to manager etcd clusters.
We are still in etcd 3.5 so we did not study the 3.6 announcement until we faced this downtime and investigated.

Anyway, even if the discovery mechanism changed, is it normal that discovery.etcd.io is down ? On the doc it says that it is not going to be maintained but does it mean it has been shut down for good ?
Is it not possible to migrate it to V3 discovery to avoid projects using it to have issues ?

[ahrtr]

have a public endpoint to create a cluster from scratch and to avoid the chicken-egg issue you get when you need an etcd cluster to manager etcd clusters.

There is no chicken-egg issue. You just need to deploy a separate etcd cluster as the discovery service to bootstrap other etcd clusters. Essentially discovery.etcd.io does exactly the same thing, it deployed the separate etcd cluster for you.

Nobody maintains discovery.etcd.io anymore.

[yogeek]

Maybe chicken-egg issue was not the right term but what I meant is that as we deploy everything with infra as code (with terraform), we should decide where to deploy this first etcd and its lifecycle as a new infra component (manage dependency with the others clusters, include it in the Disaster Recovery plan, etc.). As a new infra component, it is not just "deploy a separate etcd cluster". And that's why the publix discovery service was really useful, to avoid maintaining this little piece of infra, needed to deploy the whole platform.

What is unclear in the communication is that :

• is discovery.etcd.io officially down or is the current downtime a real issue that is going to be fixed ? Because the service is still present in the official documentation under "Public discovery service"
• "Nobody maintains discovery.etcd.io anymore." : does "no maintain" means "server is definitely shut down" ?
• finally, is a v3 public discovery service is something we can expect in the future or not at all ?

Thanks for your help