WebPagetest Forums

Full Version: Q: Acceptable use policy for recurring testing?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
AFAIK I could set up a daily test by using the REST API and a cron job on my own monitoring server.

A friend of mine has asked for something a lot like that for his cash-poor startup. What he'd like is to run a test from 5 or so locations worldwide once every 24 hours, and then chart the page load times on his internal dashboard.

I tried to look around, but didn't quite find what I was looking for. Is there a kind of acceptable use policy; would continuously running 5 to 7 tests every 24 hours be all right?

Thanks,
Yes, that would be fine. I issue API keys that have rate limits that are in the 100-200 page loads/day (each run, first or repeat view from each location is a page load).

That said, there are a lot of other free sources for ongoing performance data that he should also look at:

1 - Google Analytics - Site speed is now part of the standard reporting though it is sampled at 1% by default. You can increase the sampling rate (up to 10% right now I believe): http://analytics.blogspot.com/2011/11/si...ccess.html

2 - New Relic - They have a free tier to their monitoring suite which includes back-end and front-end performance, server monitoring and availability checking: http://newrelic.com/

3 - Browser Mob - They have a free tier that will monitor your site hourly: https://browsermob.com/website-monitoring

There may be more as well but those are the ones that I use regularly to monitor WebPagetest itself.

Thanks,

-Pat
Superb answer, thank you very much! Smile
I suggest taking a look at BrowserMob's or NewRelics 30 day trials depending on where you are optimizing. I find NewRelic is good if you are trying to optimize the application itself. Browsermob tends to provide better data for front-end/server optimization efforts.

I find that spot testing can often give you the wrong impression. I've seen metrics vary by as much as 20% within the same time (e.g. 9AM weekday) when doing ongoing testing. I've also seen 10% variance around such tests.

Also, don't get hung up on the grades and pay attention to the details. The waterfall charts are some of the most informative data you can get with webpagetest.org.

Many of the other checks are simpler optimizations for which the tool provides quick and easy way to test.
Reference URL's