Overview of historical data collection and reporting

Use historical data collection and reporting to gather useful metrics about your managed network. You can also use historical data with the chart baselining tools for predictive analysis and in situation modeling for key performance indicators.

How historical data collection and reporting works
After historical data collection has been configured and started, data samples are collected and saved in history files at either the monitoring agent or the monitoring server for the short term. The data is stored in tables, one for each attribute group for which data is being collected. If the Tivoli data warehousing facility has been set up, you can move the data to a relational database for longer term storage.

The History Collection Configuration window enables you to configure historical data collection for individual attribute groups and specify from which managed systems to collect, how frequently to collect the data, where to store it for the short-term, and whether to save data long-term to a data warehouse. You can create multiple historical collections for an attribute group, each with different collection settings and managed system distributions, and you can write filtering formulas to collect only the data samples that meet the filter criteria.

The Object Group editor enables you to combine historical data collections into groups and assign managed systems. With historical collection groups you can easily turn on and turn off historical data collection at specific managed systems or groups of managed systems by changing the distribution.

Viewing historical data
After data collection has been configured and started for an attribute group, historical data for that group can be retrieved and displayed in query-based views (chart views, table view, and the topology view).

You know that historical data collection has been enabled for the attributes in a view when you can see Time Span in the view's toolbar. The view shows current data samplings unless you use this tool to specify a broader time period. The first 24 hours of collected data is retrieved from the short-term history files; beyond 24 hours, the data is pulled from the data warehouse.

Historical data is valuable for use in chart baselining and situation modeling. Applying historical data to chart baselines and situation modeling enables you to perform predictive analysis and fine tune situations.

Requests that return large amounts of data can negatively affect the performance of monitoring servers, monitoring agents, and your network. For attributes or views that return considerable amounts of data, try to limit the time span to as short a period as will provide the required information.

Be aware that if this is a newly configured historical data collection, you will see the historical data samples that have been collected and no more. For example, if you created a historical collection and assigned it to a historical collection group an hour ago, you will see only the last 1 hour in the chart or table view even if you selected, say, the last 8 hours.

Historical navigation mode
The Set historical navigation mode tool at the end of the main toolbar is for a different purpose: It enables you to set a time span that is synchronized so that you can see the same time range in workspaces as you move among them, and compare those values.