Recently upgraded to 1.7.1 from 1.6 and have noticed an issue with the Live Visitors widget, but it may extend to other parts of piwik too.
I am tracking 750+ sites with > 100,000 page views/day. I notice when that the Live Visitors section often attempts to show (if the js doesn’t crash the browser first) a visit that contains all of the pages (or all sites, not sure) registered in the system. I haven’t been able to really figure out what’s happening here because the browser usually crashed or slows it down to the point that I can’t do anything. But I notice that loads a visit with a HUGE list of pages they “viewed”. Thousands of those little multi colored folder icons. Mousing over them shows all (or most) of the sites we are tracking.
One could be that if it see’s a GoogleBot IP on many sites it assumes it’s the same visitor and tries to display it as one. I think it should be broken down per site.
We should probably add a safety in there and cap to the first 100 pages per visit shown in the real time widget. Please, create a ticket at dev.piwik.org/trac with your report and we will fix it.
Googlebot should not be tracked by Piwik. It would be good if you could find out the row for this visitor in log_visit table by something like
SELECT * from piwik_log_visit where DATE(visit_last_action_time) = 'XYZ DATE' order by visit_total_actions desc
and put here what this visitor is. It might be some automated script running in browser scraping your website or some kind of bots. In any case, we would want to look at excluding that automatically if it’s recurrent