Enable graceful shutdown for start_{http,wsgi}_server (#999)

Signed-off-by: Antti Rasinen <antti.rasinen@relexsolutions.com>

docs/content/exporting/http/_index.md

@@ -18,6 +18,15 @@ start_http_server(8000)
 
 Visit [http://localhost:8000/](http://localhost:8000/) to view the metrics.
 
+The function will return the HTTP server and thread objects, which can be used
+to shutdown the server gracefully:
+
+```python
+server, t = start_http_server(8000)
+server.shutdown()
+t.join()
+```
+
 To add Prometheus exposition to an existing HTTP server, see the `MetricsHandler` class
 which provides a `BaseHTTPRequestHandler`. It also serves as a simple example of how
 to write a custom endpoint.

prometheus_client/exposition.py

@@ -210,7 +210,7 @@ def start_wsgi_server(
         client_capath: Optional[str] = None,
         protocol: int = ssl.PROTOCOL_TLS_SERVER,
         client_auth_required: bool = False,
-) -> None:
+) -> Tuple[WSGIServer, threading.Thread]:
     """Starts a WSGI server for prometheus metrics as a daemon thread."""
 
     class TmpServer(ThreadingWSGIServer):
@@ -226,6 +226,8 @@ class TmpServer(ThreadingWSGIServer):
     t.daemon = True
     t.start()
 
+    return httpd, t
+
 
 start_http_server = start_wsgi_server