for example, play.rust-lang.org is down for me as i write this, but when i call the ?eval or ?play commands, there's no user visible feedback that the bot tried to make an http request and failed (it gets logged as an error, but it's not communicated to the end user in any way). without a bot response, the user may think that either the bot is broken or their command syntax is wrong.
ideally, this should be done at a high level and individual commands should be able to make http requests whenever without having to explicitly handle returning a message response to the user on a timeout