Comparing Some WSGI Configurations for Plone

Lightning talk given at https://2018.ploneconf.org/

Transcript

1. Comparing Some WSGI Con�gurations for Plone Thomas Schorr
2. System Setup Vagrant VM, 2 CPUs, 2G RAM running Ubuntu 18.04 Python 3.7.1 buildout.coredev, 5.2 branch Single Plone instance using FileStorage Nginx
3. WSGI option 1: Default Waitress using 4 worker threads
4. Gunicorn op�on 1: 4 worker thread with default threading model (‘sync’) op�on 2: 4 workers using gevent worker class other possible thread classes (not tested): eventlet, gthread, tornado
5. uWSGI op�on 1: asyncio loop engine with 4 greenlet workers op�on 2: 2 preforked processes each running 2 greenlets using asyncio
6. Test runs 1000 sequen�al requests using 5 requests at a �me, 1000 requests in total � ������� �������� ����������� ����� ������ ����������� ���� � �� �� �� ���� �� � ���������������������������
9. Thank you! GitHub: tschorr t_schorr@gmx.de