I seem to be able to post multiple replies. Is there are specific thread where this happens? Are you logged in? Can you reproduce with the developer console open and see if there are any errors or warnings in the console?

Is there are specific thread where this happens?

@Esanders - I tried multiple threads. The error is consistent for me. I can reload the page and still get the same error. Also tried in Firefox and got a similar error: "RESTBase load failed: Service Unavailable".

Are you logged in?

No. The instructions don't say anything about logging in, so I didn't.

Can you reproduce with the developer console open and see if there are any errors or warnings in the console?

None other than the ones reported in the description. Note that the errors occur at load time, not at posting time.

When I try to load the page, the ext.discussionTools.parser module makes a call to RESTBase for https://en.wikipedia.beta.wmflabs.org/api/rest_v1/page/html/Talk%3ACats/407862?redirect=false&stash=true which returns a 503 server error.

It looks like the RESTBase api service just isn't running on that server. https://en.wikipedia.beta.wmflabs.org/api/rest_v1/?doc also gives me a server error.

It definitely should be (and was until recently). Not sure who is responsible for keeping those services up...

It's happening to other edits as well, so restbase is just down in general on beta, not in a talk-specific manner. E.g. https://en.wikipedia.beta.wmflabs.org/api/rest_v1/page/html/Cats/407373?redirect=false&stash=true

It definitely should be (and was until recently). Not sure who is responsible for keeping those services up...

I think @mobrovac was in charge, but he left the WMF last week. Maybe @cicalese knows.

What is the severity of this? We are looking into it. In the future, please feel free to tag Core Platform Team for RESTBase issues.

I've fixed it. A restart was needed. What happened was that metrics was configured to be sent to labmon1001 domain which apparently got reconfigured and disappeared. Puppet rerendered the config, but the service didn't get restarted and started crashing not being able to resolve a monitoring domain. This is a whole collection of bugs and misconfiguration, I'll file followup tickets.

In T239861#5757753, @Pchelolo wrote:

I've fixed it. A restart was needed. What happened was that metrics was configured to be sent to labmon1001 domain which apparently got reconfigured and disappeared. Puppet rerendered the config, but the service didn't get restarted and started crashing not being able to resolve a monitoring domain. This is a whole collection of bugs and misconfiguration, I'll file followup tickets.

Thank you for your quick help on this, @Pchelolo – the issue appears to be resolved...are you seeing the same on your end, @kaldari?

And @CCicalese_WMF, tagging Core Platform Team and communicating priority upfront...noted, we'll make sure to do both next time we encounter an issue like this.

So y'all are aware, this is/would have been a high priority issue for us as we are directing contributors to beta.wmflabs to test an early version of a prototype we intend to start iterating on in January. See: https://www.mediawiki.org/wiki/Talk_pages_project/replying/prototype_testing#Version_1.0_prototype_test