Activity tagged “performance”
A good explanation of setting up simple IP failover with Heartbeat and HAproxy.
Some tips on tuning Varnish - up to 27k req/s on a single 2.2GHz Opteron!
Comparing MySQL 4, MySQL 5, and Postgres 8. Don't pay too much attention to the numbers, but notice the shape of the graphs. MySQL does worse after about 7-10 concurrent queries, but Postgres gets the same throughput at 10 concurrent requests as it does at 100.
“PostgreSQL 8.3, by improving the Stats system, has made it very easy to find your useless indexes. These are indexes which aren't being used by any regular query, and don't enforce a constraint. Unneeded indexes cost you in several ways: they slow updates, inserts and deletes; they may keep HOT from updating the row in-place, requiring more VACUUMs; they take time to VACUUM; they add to query planning time; they take time to backup and restore. So we want to get rid of them. ”
Andrew's implementation of a “denormalized field” for Django. There's a bit I don't get about the implementation — factory functions? really? — but this is a Good Idea and ought to only get better.
Finally a good guide to tuning Postgres straight from the source.
Is Spawning Python's answer to Mongrel?
Hells yeah! I heard about this a while back, and have been hoping it would get open-sourced. Fantastic.
A cool idea: pre-bake static pages generated by Django. Sometimes simple is better.
Not sure about the quality of the benchmarks, but I'd nevertheless still like to see how Django stacks up here. Not very well I'd imagine, but Django's always been an 80% ORM anyway, so 80% of the performance would be just *fine* with me.
Loads of information about the tech behind Wikipedia. I've become convinced that the only sane way to design scalable systems is by studying the trials and tribulations of others.
Another load balancer; heard good things about this one.
Greg Mullane's complete talk on optimizing postgresql. Looks like a few things have changed since 2003 (7.3 —> 8.2), but this still seems *very* relevant.
Ridiculously in-depth explanation of what PostgreSQL means when you ask it to EXPLAIN. This is probably the only place this information lives.
Peter Van Dijck’s Guide to Ease » Blog Archive » The top 10 presentations on scaling websites: twitter, Flickr, Bloglines, Vox and more.
A *great* roundup of ten presentations on scaling: Twitter, Flickr, LiveJournal, Vox, Bloglines, last.fm, and SlideShare.
Like top, but for network usage. Highly recommended.
“Ganglia is a scalable distributed monitoring system for high-performance computing systems such as clusters and Grids.”
“All the convenience methods and syntactical sugar that makes Rails such a pleasure for coders ends up being absolutely punishing, performance-wise.” That, right there, is why Django is written the way it is.
IP load balancer with a great name.
“Varnish is a state-of-the-art, high-performance HTTP accelerator.”
Yet another benchmark of web frameworks, and yet another win for Django (this time it comes out around 3x faster than any of the alternatives).
A relatively through comparison between the speed of Symfony, Rails, and Django. Turns out Django's around 30% faster than Rails.
Contains the best definition of “Enterprise Ready” I've ever read.
Django handles lots of traffic with ease; Django sites have survived slashdottings, farkings, and more. Here are some notes on how we tweak our servers to get that type of high performance.