Over the past few years I have become a crusader in using tools such Piwik Analytics in Blackboard Learn. In many presentations I have shown how these tools have impact how we support the Blackboard application at the University of Missouri. Many times I get emails, chats, or questions at conferences about Piwik and Blackboard Learn. So I’ve decided to put fingers to keyboard and answer some of the most often asked questions I get about the subject.
Question: In trying to size the storage for the system, how much data are you collecting in a typical month, year, or whatever data point you might have easily accessible? How long are you keeping the data?
I keep about 180 days of visitor logs and after that time Piwik automatically removes the data for me. Piwik retains the statistical data using reports which provide daily / hourly statistical information, but the raw logs are removed. I currently use about 13 GB of database storage, you can see the breakdown in the Piwik slideshow below.
Question: How can I understand the amount of load Piwik will need to be able to handle?
A great way to understand your current load is to look within the database of your Blackboard Learn environment. Blackboard keeps a daily record of many items in the SYSTEM_REPORTING table. the regularly see 1.3 million page views a day and about 300k course views so our environments are comparative.
select timestamp, page_views, course_page_views
from bb_bb60.system_tracking
where timestamp > sysdate – 30
order by timestamp desc;
Question: How did you install Piwik?
Up to date directions are available at: http://piwik.org/faq/how-to-install/
Question: What are your database and web server configurations?
Both servers are virtualized. Here are the specs of each.
Piwik Database – Intel(R) Xeon(R) CPU E7- 4850 @ 2.00GHz Dual Core processor, 16GB RAM, and 63GB of storage for the database, and approximately 25GB for OS and temp.
Piwik Application – Intel(R) Xeon(R) CPU E7- 4850 @ 2.00GHz Dual Core processor, 8GB RAM, and 15GB of storage for application, OS, and temp
Question: Have you ever experienced issues that would be related to Piwik?
I haven’t seen any performance issues with the Piwik application itself. When starting this project Piwik used shared database hosting and the size and I/O put quite a bit of load on those resources. Recently the Piwik database moved to its own database server and performance has improved since. Now when you do have to do maintenance, you will need to make sure to put the server in maintenance mode and if you have to restart the database without it in maintenance mode you can create some major issues. But maintenance mode is pretty easy to turn on, see http://piwik.org/faq/how-to/#faq_111
What do you do with that data? How does it impact your decisions?
Your Blackboard ate my homework will really be solved by searching the logs and the activity accumulator at this time. Piwik really helps you better understand how your instance is used such as peak times, OS and browser usage, locations, search tools, performance, etc. We have used it to find out how a browser issue will impact users or help us develop testing plans even plan mobile strategies.