Anyone here ever add hardcoded sentry environment names, no I don't mean hardcoded names, limit the allowed ones? (e.g. not "dev" but "Dev")

Am I the silly one that realized that this is probably not a good way of handling it and created a PR to remove the check?

No, I didn't. I don't have the server hardware anymore to run the sentry these days. 8gb if ram? I removed all memory checks to just run it. I couldn't believe it. And yes, whole server crashed when running.

Having sentry is almost a flex.

@retoor Didn't realize they allowed self-hosting, guessing I got another reason to work on that NAS. Those are some high demands for something that's basically a glorified HTTP server… (for web dev it's super easy to implement your own error manager)

@BordedDev I created my own central logger after this: https://molodetz.nl/retoor/mololog but it misses two things: application name and the exception itself. Only message is supported. Logging is done by a thread pool and it will not cost performance to the logging client so it's ok to do massive amount of log lines. The exception and name is not hard to implement if you're interested. The patch function patches all loggers. In the end, it's just a traditional python native logger. No magic here. It doesn't have a frontend. Don't know if I will still work on it. Working on visit stats instead

When I started to use sentry, it was in functionality already huge but installation was just pip install and setting up a postgress server. It was a bit before docker became popular. But in resources, it was low as a logging service should be. I have no idea how such server could need 8gb. Would say total opposite of few mb.

But yh, sentry is cool stuff and indeed also self hosted.