Anthony

I believe https://github.com/readthedocs/readthedocs.org/issues/11061 is replicating this behavior in production, but waiting to confirm. If confirmed, we should probably get a fix for this issue on the roadmap.

Noted here: - https://github.com/readthedocs/readthedocs.org/issues/11061#issuecomment-1918352645 The PR fixing 11061 confirms (to me) that this bug is affecting production as well. At very least the beta instance is affected by the bug....

Your results sounds about right for reproducing locally. What does the page response time look like without errors? In production, as the database is larger, response times from this are...

Nice! I had tried a similar fix that you suggested ([here](https://github.com/readthedocs/readthedocs.org/issues/10954#issuecomment-1856237283) though the permalink looks off), but it didn't help much. Maybe your fix above affects things at a deeper...

This can be tested further on web-extra with a template change and dropping the settings for sentry et al

> I attempt to uninstall it and re-installing it, but I broke something and the instance started using the old dashboard Yeah, I've encountered this, the package is not installed...

I just tested this and it worked. :shrug: I was able to edit the base template and the changes showed in the beta instance.

> I may not be sure what I'm looking for here, then. With "it didn't work" I meant that I still got Bad gateway Error code 502 I think I'm...

Just to clarify, when you say 502, are you noticing a long response time too? I think I debugged the build list issue with pyinstrument in production, or if not,...

I just opened #46 which may dictate what we do here. I agree we should have separate docs if we keep this logic separated from RTD though. 👍