Adventures in TDW land

Tivoli Data Warehouse (TDW) is a very useful and powerful feature of the OMEGAMON suite and of Tivoli monitoring in general.  Each Tivoli monitoring solution from Linux, UNIX, Windows to z/OS connects to the Tivoli infrastructure and may optionally send information to the Tivoli Data Warehouse.

When you enable TDW history collection you specify many options, including what tables of information to collect, how often to collect, what agent types and managed systems to collect from, and if summarization/pruning is required.  While seemingly straightforward, each of these options has important considerations that may impact the usefulness of the resultant data being collected.

With many users there can be quite a few questions.  Where to begin?  What data should be collected?  How should the data be retained and for how long?  Where should the data be stored?  How is the data to be used and by what audiences?

Planning and analysis is important to a successful implementation of the TDW.  One approach that should be avoided is the 'turn it all on' strategy.  The turn it all on approach will inevitably result in the user collecting more data than is needed and this has multiple shortcomings.  First, unnecessary data collection wastes space and resources.  Second, unnecessary data collection makes it slower and more time consuming to retrieve information that is useful. 

As a general methodology, it is usually better to employ a start small, then work your way up approach to enabling TDW history collection.  You can always dynamically enable more collection options, but weeding out large quantities of useless data may be a time consuming exercise.

I will be doing a series of posts on TDW with the goal of documenting a best practices approach to enabling this portion of the tool.