Profiling Mondrian MDX requests in a production environment

Raimonds Simanovskis from eazyBI
Raimonds Simanovskis from eazyBI

When working with Pentaho Mondrian server it can happen to you that some MDX requests are slow. To identify the problem, you can activate Mondrian debugging options. However, if you have many users and schemas working on Mondrian, this is not an option. Raimonds Simanovskis will demonstrate at Pentaho Community Meeting how to debug MDX request problems in a Mondrian production environment.

Who are you?
I am Raimonds Simanovskis, the founder and CEO of the business intelligence web application eazyBI. Previously, I had a long “enterprise software development” history but later I became more interested in open source technologies. I mostly use (and love) Ruby and JavaScript programming languages and have several Ruby open source projects (including mondrian-olap which wraps Mondrian engine in a JRuby library). I am also an open data activist and have organized and mentored several open data hackathons in Latvia. I am a frequent speaker at conferences and other events and like to share my experience about software development and bootstrapping a business still keeping the balance of being a half geek and half entrepreneur.

What will your talk be about? Why is the topic relevant for (Pentaho) users?
I will talk about profiling Mondrian MDX requests in a production environment. In our eazyBI cloud environment, we have thousands of customers that we serve with one single application that embeds Pentaho Mondrian OLAP engine. When some MDX requests are slow, then in a development single user environment it is easier to turn on all Mondrian debugging options and investigate what is happening. But in a production environment with thousands of active schemas and users, it is much harder to isolate and debug why some specific requests are slow.

Collecting MDX profiling information in Mondrian
Collecting MDX profiling information

In the presentation, the available Mondrian profiling and logging options will be described and how we use them in our production environment. I hope that these solutions and ideas can help other Pentaho Mondrian users who would like to have more profiling and debugging options in their production environments where Mondrian is used.

What do you expect from PCM?
This will be already my fourth Pentaho Community Meeting, previously I have been twice in Antwerp and last year in Mainz. As we are using just a small part of Pentaho solutions it is interesting to learn about different other use cases and how Pentaho is used for business intelligence projects. It is nice to meet Pentaho community members as they are always friendly and ready to share their experience. And I also like visiting Italy and this will be my first time in Bologna.


Liked this article? Then check out these: