Hi, On Wed, 2004-09-29 at 10:32, John Poltorak wrote:
On Wed, Sep 29, 2004 at 10:28:57AM +0200, Max M wrote:
John Poltorak wrote:
How do people count the number of hits they have on a Zope website?
You are inprecise here. Please define "count hits"
Do you mean a simplehit counter for one or more pages, or do you mean complete statistics?
I'm none too clear myself, I hoped people would understand what is meant by a 'hit'. Websites often say they have so many 'hits' per day. I mean whatever they mean.
Well, like you, there is basically nobody really knowing "whatever that means". No matter if they are still constantly speaking about this. Remember: this is marketing So you have 2 kind of figures, where one is precise and has a meaning and the other is mostly rubbish (depends on your setup) One figure is how many and how often how many bytes are loaded off your webserver. This is in the access logs. Either Z2.log or for example the apache log if you use apache in front of your zope - as most ppl. do. This tells you used bandwidth, peaks, "not found" and other status information, how much cached responses and so on. You can use this figure to plan your server hardware/software. The other figure is the vague imagination of having people out there, physically represented by their computers which browse thru your site and load a page with embedded objects each time so you have kind of a trail and perhaps see how many times a given page is "viewed". But: you dont know. You never know. Unless you have a very strong session management where you forbid and inhibit any caching and let only known users get in and _never_ cache anything. Then you would have a chance of exact numbers as of which elements were loaded - but you dont know if the user actually looked at it :-) There are in fact a lot of caches - and you want them to have nice performance - so the 2nd figure is far from exact in any way. If marketing people insists of having such numbers use any logfilefaker script you find appropriate - the faked numbers are as good as any "believed to be measured" ones. Regards Tino