NoneType object has no attribute put http meta

[Ismael]

Using sdk version 2.4.2

Traceback (most recent call last):
  File "/usr/local/lib/python2.7/dist-packages/django/core/handlers/exception.py", line 41, in inner
    response = get_response(request)
  File "/usr/local/lib/python2.7/dist-packages/django/core/handlers/base.py", line 187, in _get_response
    response = self.process_exception_by_middleware(e, request)
  File "/usr/local/lib/python2.7/dist-packages/django/core/handlers/base.py", line 227, in process_exception_by_middleware
    response = middleware_method(request, exception)
  File "/usr/local/lib/python2.7/dist-packages/aws_xray_sdk/ext/django/middleware.py", line 102, in process_exception
    segment.put_http_meta(http.STATUS, 500)
AttributeError: 'NoneType' object has no attribute 'put_http_meta'

[chanchiem]

Hey,

Can you give a code snippet of what's triggering this? Is this ran on Lambda? I need a little bit more information so that I can try reproducing this and isolating the cause. By any chance, is this ran using the Serverless model? Thanks in advance.

[Ismael]

I'm doing this in urls.py, to patch all libraries

from aws_xray_sdk.core import xray_recorder, patch_all
xray_recorder.begin_segment('startup')
patch_all()
xray_recorder.end_segment()

This is my config

XRAY_RECORDER = {
    'AUTO_INSTRUMENT': True,
    'AWS_XRAY_CONTEXT_MISSING': 'LOG_ERROR',
    'AWS_XRAY_TRACING_NAME': 'nTitleMe',
    'PLUGINS': ('ECSPlugin',),
}

This is run locally.

I can reproduce this with even a simple view that only raises an exception

def aview(request):
  raise Exception()

I have lots of custom middlewares, but XRay's is first.

[chanchiem]

Everything looks good with your configuration. What version of Django are you using?

We currently support only the following versions:
django >= 1.10, <2.0

[Ismael]

I've got: Django==1.11.15

…

On Wed, 3 Apr 2019 at 14:35, Chan Chiem Jeffery Saeteurn < ***@***.***> wrote: Everything looks good with your configuration. What version of Django are you using? We currently support only the following versions: django >= 1.10, <2.0 — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#152 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAHVzsX0228chOWRpsD9qSkR6I6-Cj8Gks5vdOZugaJpZM4cYsR7> .

[chanchiem]

Is it possible for you to temporarily disable the other middlewares and see if this issue still persists? I'm thinking that maybe there's a middleware that's causing this issue. If the issue does goes away, can you also isolate which middleware is causing this exception? I still can't seem to reproduce this issue you're having on my local system, which is leading me to think that maybeone of the middlewares is causing issues with X-Ray.

I also notice that you're using an ECS plugin. Is this plugin being used with X-Ray when you develop in your local development as well? Just as a sanity check, does removing the plugin from the configuration help?

[aarcro]

I'm getting this on Runsever as well, Django 2.2.3

[aarcro]

I rolled Django Rest Framework back to 3.9.4 And this went away. 3.10 was not working

[levesquejf]

I just had the same issue with flask 1.1.1 and aws-xray-sdk 2.4.2 running on AWS Lambda (Python 3.7) and deployed with Zappa. Everything (pynamodb, botocore, requests) is patched with patch_all(). I have not been able to reproduce.

File "/var/task/flask/app.py", line 2446, in wsgi_app
    response = self.full_dispatch_request()
  File "/var/task/flask/app.py", line 1952, in full_dispatch_request
    return self.finalize_request(rv)
  File "/var/task/flask/app.py", line 1969, in finalize_request
    response = self.process_response(response)
  File "/var/task/flask/app.py", line 2266, in process_response
    response = handler(response)
  File "/var/task/aws_xray_sdk/ext/flask/middleware.py", line 74, in _after_request
    segment.put_http_meta(http.STATUS, response.status_code)
AttributeError: 'NoneType' object has no attribute 'put_http_meta'

[OJFord]

quart==0.18.4, flask==2.2.5, aws-xray-sdk==2.14.0

Traceback (most recent call last):
  File "/var/task/quart/app.py", line 1735, in finalize_request
    response = await self.process_response(response, request_context)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/var/task/quart/app.py", line 1762, in process_response
    response = await self.ensure_async(function)(response)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/var/task/quart/utils.py", line 61, in _wrapper
    result = await loop.run_in_executor(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/var/lang/lib/python3.11/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/var/task/aws_xray_sdk/ext/flask/middleware.py", line 74, in _after_request
    segment.put_http_meta(http.STATUS, response.status_code)
    ^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'put_http_meta'

Even if unable to reproduce, perhaps we could just fail gracefully if no segment? i.e. best effort to set annotations but not error?

[jcampbell05]

This appears to happen after "Subsegment api discarded due to Lambda worker still initializing" they could be related ?

[jcampbell05]

It looks like for whatever reason the flask/django integrations fail to start a segement which causes the problem, especially when lambda is still setting up.

I had to add this code xray_recorder.begin_segment("DefaultSegment", traceid=None) and it no longer crashes, however I do not think this is the correct approach.