Send timeout triggered by filtered events

[lobsterkatie]

Currently, if an event triggers any of Sentry's server-side filters (like the web crawler filter, for instance), it kicks the ruby SDK into timeout mode, because a filtering error is treated the same way as, say, a rate limit error. Future legit errors therefore get blocked. Compounding the problem, the timeout never ends unless the SDK is entirely idle for 36 seconds. (See the related issue here.)

Other SDK's don't go to timeout mode in response to filtered events. Let's make this configurable, at minimum, or take out the behavior entirely.

[st0012]

@lobsterkatie can you try 3.0.x versions to see if this is still an issue? I think https://github.com/getsentry/raven-ruby/pull/886 has fixed the issue.