PyconAU 2012: Server performance

Notes from a talk about server performance at Pycon AU 2012. Posted by Thomas Sutton on August 19, 2012

The low-hanging fruit when improving perceived performance is on the front-end, not the back-end, and certainly not the small overhead of the HTTP server.


processes = 2-4 * cpus

The GIL essentially serialises thread processing. This impacts throughput in threaded configurations.

Reverse proxy with nginx, etc. to decouple clients and app servers. Need to be able to kill backlogged requests when they aren’t needed; restarting some servers won’t clear a backlog.

Autoscaling configurations can cause load problems with fat applications; pre-configuring for required maximums.

Server monitoring tools largely treat web apps as a black box, they show the effects, not causes.


New Relic.

Apache modules are like “batteries included” for web-servers.

This post was published on August 19, 2012 and last modified on June 4, 2019. It is tagged with: event, PyconAU 2012, Python, performance.