We now have a pretty good health-check job that runs regularly and checks a lot of things.
Right now it reports to Sentry (as it should) but does not report in the "correct" format so that means it's hard for us to identify issues. Here are improvements that need to be made:
[x] We should report "individual" issues to sentry.
[x] Distinguish between "warnings" and "error" (warning are non critical, such as "misconfigired token" or "pending proposal on DAO", while errors are critical, like "bad version"... etc
We now have a pretty good health-check job that runs regularly and checks a lot of things. Right now it reports to Sentry (as it should) but does not report in the "correct" format so that means it's hard for us to identify issues. Here are improvements that need to be made: