We're making several major changes to the daemon's functionality in the 0.3 release, so we'll need to establish a performance baseline to make sure we're aware of any slowdowns that are introduced by new features. Also, as we deploy this at LAIKA, we'll almost certainly start to exercise it in ways we haven't anticipated, so we'll want to be able to gauge the effectiveness of our optimizations.
We've investigated several different HTTP benchmarking tools, and the ones that look like they'll be most useful to us are:
Some of the features that we'll want:
- Historical plots of performance vs. svn rev
- A way to cross-reference performance and profiling output
- In-browser benchmarking
