asked by management to give page response times for a few key pages
"page response times" Is that server response time, download time, page load until ready time, responsiveness after the page is loaded and ready, or something else?
These two tools might be useful in identifying some bottlenecks, but no personal experience with them, and they really only look at the client end of things; not what the server is specificlly dealing with. I took only a cursory look at the sites, but didnt see anything about benchmarks. They are Firefox plugins.
On the server, you can enable tracing as a simple way to get specific relative timing, but it adds it's own overhead so the times will be slower/longer than on production without tracing.
The metrics I would provide stackeholders are the ones that you are required to provide, or those that the users are particularly interested in. If they are interested in page response times, I would give them benchmarks and ask them for threshholds. In otherwords, this is what it is, what is acceptable to you? What is your reasonable goal? Perhaps what they want or need is not possible in a single web action and single view given strict thresholds, and it will have to be partitioned to multiple actions and multiple views.
The benchmarks I would give them, especially when they are concerned with a few specific pages would include:
- the pages they are interested in
- a sample of other dissimilar, less complex pages. The reason I would include a sample of other dissimilar, less complex pages is to show that there is a certain overhead simply due to hardware infrastructure and software stack that only improves by primarily more investment into those resources.
- a sample of other pages that represent similar complexity or content weight as the pages they are intersted in. Perhaps there is room for refactoring, or optimization in the pages of interest and this other similar sample might demonstrate it.
