Count process as ready if it can respond

Our 'readyness' condition check is to make sure we can show
something to the user - it doesn't matter if it is a non-200 response.
Primarily, we were getting 404s and other responses that were causing
issues here - 3xxs were resolved by aiohttp.

Author: yuvipanda

docs/server-process.rst

@@ -41,6 +41,9 @@ pairs.
 
    Timeout in seconds for the process to become ready, default ``5s``.
 
+   A process is considered 'ready' when it can return a valid HTTP response on the
+   port it is supposed to start at.
+
 #. **environment**
 
    One of:

jupyter_server_proxy/handlers.py

@@ -298,8 +298,10 @@ async def _http_ready_func(self, p):
         async with aiohttp.ClientSession() as session:
             try:
                 async with session.get(url) as resp:
+                    # We only care if we get back *any* response, not just 200
+                    # If there's an error response, that can be shown directly to the user
                     self.log.debug('Got code {} back from {}'.format(resp.status, url))
-                    return resp.status == 200
+                    return True
             except aiohttp.ClientConnectionError:
                 self.log.debug('Connection to {} refused'.format(url))
                 return False