fix: increase the gunicorn timeout to 40 seconds

For some reason, when we keep the default of 30 seconds, request get
killed after 20-25s. Not sure why, but --timeout=40 fixes the problem.
We still never reach 30seconds, because we now have a 25s budget on the
g2p conversion, which is the most expensive step.

Procfile

@@ -1,2 +1,2 @@
 # Command for launching the web API server for ReadAlongs-Studio on Heroku
-web: gunicorn --workers 5 --worker-class uvicorn.workers.UvicornWorker readalongs.web_api:web_api_app
+web: gunicorn --workers 5 --worker-class uvicorn.workers.UvicornWorker --timeout 40 readalongs.web_api:web_api_app