Deployment tooling

[saethlin]

I'm running this with an S3 bucket, and a domain I control hooked up to a Cloudfront distribution, and I mostly actually run the software on my desktop or a spot instance.

I feel like it should be possible to at least automate setting up the instance.

[saethlin]

I've adjusted the path computations in the web pages, so this can be served straight out of an S3 bucket without issue. Maybe that alleviates any desire to have orchestration for a Cloudfront distribution?

[saethlin]

I've been mulling over sticking this whole thing into a database to make it easier to query, but the internet says that for such huge chunks of what is just binary data, that the filesystem is as good as any database. Hm. Maybe that means I should be querying and uploading to a tiny VM with a large attached disk.