We want to prevent our healthcheck endpoints from reporting to BugSnag if they raise an error. Since these endpoints are often being triggered at a high frequency, if they report to BugSnag everytime, we will quickly exceed our BugSnag quotas. Ultimately, the healthcheck probe will be aware of the errors anyway, so BugSnag doesn't need notified. This was the solution we came up with for catalog:
Bugsnag.configure do |config|
config.add_on_error(proc do |event|
action = event.request&.dig(:railsAction)
event.ignore! if action&.start_with?('ok_computer')
end)
end
Note that this exact configuration might not work for this application. However, if all healthchecks are ran through OkComputer, then it should work.