Close segment in only _handle_exception in case of Internal Server Error

[srprash]

Issue #, if available: #269

Description of changes:

• Return from _after_request when response code >= 500 so that the exception in recorded on the segment in the _handle_exception method.
• Not enabling the testing mode so that the request is handled by the application in usual way during the exception

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

[codecov-io]

Codecov Report

Merging #271 (1f04fd4) into master (5af2357) will decrease coverage by 0.07%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #271      +/-   ##
==========================================
- Coverage   79.46%   79.38%   -0.08%     
==========================================
  Files          82       82              
  Lines        3253     3255       +2     
==========================================
- Hits         2585     2584       -1     
- Misses        668      671       +3     

Impacted Files	Coverage Δ
aws_xray_sdk/ext/flask/middleware.py	88.75% <100.00%> (-3.56%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5af2357...1f04fd4. Read the comment docs.

[willarmiros]

LGTM

[ianmetcalf]

@srprash I think this introduces a subtle bug. A 500 response in after_request doesn't always result in an exception being passed to teardown_request. For instance the following would cause the segment not to be closed at all.

app.route('/error')
def error():
    return 'Server Error', 500

[srprash]

@ianmetcalf Ah! I see. I was kind of unaware of tuple as a flask application response. I just found out about it here.
Good catch! I guess we should always close the segment in teardown_request if one is still present. I'll open a PR for the fix shortly.