BugFix: Always close segment in teardown_request handler

[srprash]

Issue #, if available: #269

Description of changes:
The previous PR which intended to fix the #269 introduced a bug where if the flask application does not pass an exception in case of response code >= 500, the segment will remain unclosed. This may be an edge case scenario that application faults do not pass exception to teardown_request handler but it is still a valid case of responses.
Thus, irrespective of exception being present or not, the middleware should always end the segment if its present.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[srprash]

@ianmetcalf Feel free to take a look at this.

[ianmetcalf]

@srprash This should take care of it.

With this change in place could you skip closing the segment in after_request for all requests. Having teardown_request close all segments would make things easier to reason about.

It might also be a good idea to rename the _handle_exception method since it's not only for exception handling now.

[srprash]

@ianmetcalf That's a fair call out since the teardown_request is bound to be called at the end of each request, we can avoid some redundant code for closing the segment. And +1 for renaming the method.

[willarmiros]

LGTM