[loxs]

You can migrate to the heavier infra later. I am serving hundreds of concurrent users from a single Rust binary backed by Sqlite. So far it hasn't shown the slightest problems and migration can happen if the service grows an order of magnitude (or two) from here.

[emacsen]

Migration/upgrades need to happen no matter the size of the service, so I'm not sure what you mean.

[loxs]

Sorry, I mean porting to more flexible infra, not DB migrations. For these I use downtime right now... it's not a big deal.