0

I have written a set of python REST APIs that are served by the Tornado web framework. The issue I am facing is like this: When handling endpoint1 or API1, I need to get at some data that endpoint2 or API2 can provide. So, inside the handler for endpoint1, I call something like this:

   class endpoint1(tornado.web.RequestHandler):
    .........   

      def get(self):
      ..........
         http_client = AsyncHTTPClient()
         url = "http://127.0.0.1:8686/v1/endpoint2"
         response = yield http_client.fetch(url)

But, the code just hangs at this point. My guess is it doesn't work since the framework is currently in the middle of servicing endpoint1 and I am trying to sneak in another request within. I am looking for suggestions on how to get this to work without using MQ or databases. I tried using nest_asyncio also - no dice. Any help appreciated

Nagendra Murti
  • 33
  • 1
  • 4
  • 1
    The code example you've posted works fine (i.e. one handler can make http request to another handler). There must be some other problem in your actual code. – xyres Jan 08 '21 at 20:45
  • or run tornado behind loadbalancer and have it make a request out to the LB, that way you hit a different tornado process. yes it requires more integration but scales better anyway. – cowbert Feb 03 '21 at 13:27

1 Answers1

0

Turns out that nest_asyncio actually does the trick. Here is a link to another thread that explains it well: RuntimeError: This event loop is already running in python

import nest_asyncio
nest_asyncio.apply()

 class endpoint1(tornado.web.RequestHandler):
.........   

  def get(self):
  ..........
     http_client = AsyncHTTPClient()
     url = "http://127.0.0.1:8686/v1/endpoint2"
     response = await http_client.fetch(url)
Nagendra Murti
  • 33
  • 1
  • 4