Thomas Sibley

Thanks for this prototyping, @jameshadfield. I think at the very least we should be doing some form of server-side rendering: either ahead-of-time (SSG) or on-request (traditional SSR) or both. The...

Option 2 seems nicer to me. It gives us precise control of the build/deploy process and sets us up better for future deploy improvements (or even someday moving off Heroku)....

The outstanding PR https://github.com/nextstrain/nextstrain.org/pull/141 from a while ago is also relevant, although got tangled up in caching issues and I never got back to it.

> Is there an agreed standard for general want-to-always-see-this logging when using the 'debug' module? Its [docs](https://github.com/debug-js/debug?tab=readme-ov-file#usage) seem to suggest we'd do something like `debug('nextstrain')` and then habitually run with...

The issue is that we haven't considered what it means to `PUT` and `DELETE` past versions (and how that affects `OPTIONS`). We can't support replacing past versions with `PUT`, so...

Ah, and this is also a place where having a `ResourceVersion` class in the sources data model might be more helpful (compared to a version-aware `Resource`): to let us more...

This is part of the larger umbrella of UI for Groups. https://github.com/nextstrain/private/issues/57

I've noticed this and believe it's due to how Heroku's routing layer switches things over a bit early when cutting between the old dynos and new dynos. I wouldn't call...

This would be nice, for sure. It's totally possible. I not-infrequently run in true production mode with local Redis, local SSL, local secrets and encryption keys, etc. But yeah, it's...

Some background: I held off (for years) on introducing something like Terraform because it is additional complexity. But eventually I really couldn't justify holding off any more. I needed to...