
In Go it's really easy to integrate BugSnag with the logrus library. We can then hook up Bugsnag to Pagerduty to notify our on call if there is a sudden spike in errors. Review collected by and hosted on G2.com.
The information Bugsnag captures in Go isn't that useful. Stack traces can be misleading. I'd like to track multiple levels of log in Bugsnag but not necessarily report them all through to Pagerduty (i.e. I'd like to also capture warnings but not have them ping the on call). It's very much an all or nothing tool that can only be configured at a coarse level.
Additionally the options for when it should notify aren't that useful. The option for exponential backoff on the notifications seem to be "for all time" whereas it would be more useful if it reset after a short period. If the errors settled down and we didn't see any for a week, I want a single new instance of the same error to notify me again. Review collected by and hosted on G2.com.