I plan to install Piwik on a large and special architecture, that is composed of several servers (having their own Apache and MySql database) for load-balancing and disponibility purposes.
But i’m not sure Piwik will allow us to collect request separately as we usualy do, then collect and gather datas regularly.
I’m seen this ticket (but closed) about implementing sharding in Piwik code
http://dev.piwik.org/trac/ticket/386[/url], and this development try: [url=http://dev.piwik.org/trac/ticket/471]http://dev.piwik.org/trac/ticket/471, but its doesn’t seem to be good solutions.
I’m looking at the bulk load feature currently in development: http://dev.piwik.org/trac/ticket/134
For the present, it’s maybe a good way to proceed: allways storing “piwik logs” from several servers, then processing all of them in one database.
Can you share me your opinion on all this. Maybe i’m missing some nice feature ? If no, do you think that the best solution is to wait for the previous development ?