Log unknown errors from the client somewhere on the backend

[jmoggr]

It can be useful to have client side errors logged like any of the other services. This can allow for capturing more detailed errors compared to what users might provide in a ticket, and it also gives us a heads up that users are experiencing errors. It also allows us to capture all errors, instead of just the ones users create tickets for.

There are 3rd party services for helping to manage the instrumentation and logging like sentry.io, or you can just create a new api endpoint and dump the errors in a table somewhere.

This should take less than a day, and really depends on how fancy we want to get with instrumentation and monitoring/alerting.

[epatters]

Good point. Our original, prototype backend, written in Node/TypeScript, had sentry.io running to log errors, but that got lost along the way when we rewrote the backend in Rust. Having an integrated setup to log errors on both the frontend and the backend will surely pay for itself many times over.