gRPC support

[Private-SO]

Describe the bug

I have a aws alb load balancer that runs on port 443 with a self signed certificate on it. Behind alb there is a target group(ec2) that runs on GRPC protocol version. The target group(EC2 instance) has nginx listening on port 50052. When i trigger a HTTP2 POST request to my alb with verify=False flag I am getting a random error. Sometimes I get the correct response which is a string and some times I get a error.

To reproduce

1. Create a application load balancer with GRPC protocol_version target group(ec2 instance).
2. SSH into the ec2 instance and install nginx and change the configuration to create a simple location like this

My nginx configuration

server {
        listen 50052 http2;
        location =/version {
        if ($request_method = POST) {
        return 200 '{"field":"value"}';
}
}
}

3. Allow the alb security groups and ec2 security groups to allow the outside traffic.
4. SO now use httpx to trigger a POST request. below is my httpx code.

import httpx
r = httpx.Client(http2=True,verify=False)
response = r.post('https://alb-913799985.us-east-1.elb.amazonaws.com/version')
print(response.text)

Expected behavior

{"field":"value"}

Debugging material

Traceback (most recent call last):
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_exceptions.py", line 326, in map_exceptions
yield
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 856, in _send_single_request
(status_code, headers, stream, ext) = transport.request(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_transports\default.py", line 102, in request
return self._pool.request(method, url, headers=headers, stream=stream, ext=ext)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\connection_pool.py", line 218, in request
response = connection.request(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\connection.py", line 106, in request
return self.connection.request(method, url, headers, stream, ext)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\http2.py", line 119, in request
return h2_stream.request(method, url, headers, stream, ext)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\http2.py", line 297, in request
status_code, headers = self.receive_response(timeout)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\http2.py", line 371, in receive_response
event = self.connection.wait_for_event(self.stream_id, timeout)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\http2.py", line 198, in wait_for_event
self.receive_events(timeout)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpcore_sync\http2.py", line 215, in receive_events
raise RemoteProtocolError(event)
httpcore.RemoteProtocolError: <StreamReset stream_id:1, error_code:ErrorCodes.INTERNAL_ERROR, remote_reset:True>

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "D:\pipenv\http2request.py", line 29, in
response = r.post('https://alb-913799985.us-east-1.elb.amazonaws.com/version')
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 987, in post
return self.request(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 728, in request
return self.send(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 762, in send
response = self._send_handling_auth(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 800, in _send_handling_auth
response = self._send_handling_redirects(
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 832, in _send_handling_redirects
response = self._send_single_request(request, timeout)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_client.py", line 856, in _send_single_request
(status_code, headers, stream, ext) = transport.request(
File "D:\python3.9\lib\contextlib.py", line 135, in exit
self.gen.throw(type, value, traceback)
File "C:\Users\rocki.virtualenvs\pipenv-LAdtM08T\lib\site-packages\httpx_exceptions.py", line 343, in map_exceptions
raise mapped_exc(message, **kwargs) from exc # type: ignore
httpx.RemoteProtocolError: <StreamReset stream_id:1, error_code:ErrorCodes.INTERNAL_ERROR, remote_reset:True>

Environment

• OS: Windows
• Python version: 3.9.2
• HTTPX version: 0.17

[florimondmanca]

@Private-SO Hey,

If you'd like folks to give a helping hand, could you please follow and fill in the issue template? There's important information there that needs communicating anyway to troubleshoot situations like these. Thank you!

[Private-SO]

I have updated the details regarding my issue. let me know if you need any further information. Thanks

[lovelydinosaur]

Hrm... it's a bit tricky to replicate this as:

>>> response = r.post('https://aws-xxx.us-east-1.elb.amazonaws.com/version')
...
httpx.ConnectError: [Errno 8] nodename nor servname provided, or not known

[Private-SO]

sorry, I didn't clearly described my issue. I have updated my issue and also provided my ALB URL. I am sure it is enough to reproduce this from your side. Let me know if you need any further details. Thanks

[lovelydinosaur]

Create a application load balancer with GRPC protocol_version target group(ec2 instance).

This is way outside what we're able to provide our time to help with. I wouldn't expect to necessarily be able to make gRPC requests, because at least as far as I recall it doesn't quite function like standard HTTP/2, but I'd need to spend a whole heap of time getting my head back into what we'd need if we did want to support gRPC.

[lovelydinosaur]

Some brief notes related to this at #1150 and #1274