IBM Tivoli Monitoring > Version 6.3 Fix Pack 2 > Administrator's Guide > Manage historical data > Summarization and pruning configuration
IBM Tivoli Monitoring, Version 6.3 Fix Pack 2
Best practices for summarization and pruning
Use a best practices approach in determining how to summarize and prune the data samples stored in the Tivoli Data Warehouse.
Before enabling historical collection think about your business requirement for the data. There are four common use cases for the historical data. Your needs will vary for each attribute group, so consider the use cases when configuring historical collection: problem determination and debugging; reporting; capacity planning and predictive alerting; and adaptive monitoring.
For performance tuning best practices for the Summarization and Pruning agent, as well as the other monitoring components, see "Performance tuning" in the IBM Tivoli Monitoring Installation and Setup Guide.
Each of these use cases has different historical requirements. The following sections describe each of these use cases and the types of historical collection that will be desirable.
- Problem determination and debugging
- These types of metrics are used for problem determination and debugging, which tends to be relatively short term in nature. Occasionally there is a need to compare performance from a long time ago, but most of the time Subject Matter Experts (SMEs) want to go back a few days and evaluate the performance of a server or application and compare it to the current performance. In this case, there is no need for summarization of the data.
- Reporting
- When configuring historical collection, consider the purpose of your reports. Some reports are used for long term trend analysis, some reports are used to show that SLAs are being met, and some reports are relatively short term to show the health of a server. The primary driver of the historical collection is the duration of the reports. For short term reports, you can use Detailed data. For short to medium term reports, use Hourly summarized data. For medium to long reports use Daily or Weekly summarization.
- Keep in mind when configuring summarization, that you do not need to configure all intervals. For example, if you want Weekly summarization, you do not need to also configure both Daily and Hourly. Each summarization interval can be configured independently.
- For information on using the Summarization and Pruning agent to create, populate, and maintain the dimension tables required for Tivoli Common Reporting, see Create and maintain the dimension tables.
- Capacity planning and predictive alerting
- For capacity planning and predictive analytics, you typically perform long term trend analysis. The Performance Analyzer, for example, uses Daily summarization data for the predefined analytic functions. So, in most cases, configure daily summarization. You can define your own analytic functions and use Hourly or Weekly summarization data.
- For the analytic functions to perform well, ensure that you have an appropriate number of data points in the summarized table. If there are too few, the statistical analysis will not be very accurate. You will probably want at least 25 to 50 data points. To achieve 50 data points using Daily summarization, you must keep the data for 50 days before pruning. More data points will make the statistical predictions more accurate, but will affect the performance of your reporting and statistical analysis. Consider having no more than a few hundred data points per resource being evaluated. If you use Hourly summarization, you get 336 data points every 2 weeks.
- Adaptive monitoring (dynamic thresholding)
- The situation override capability enables you to analyze historical data to define a threshold that is based on the past performance characteristics. You can define time of day and shifts to analyze the historical data and make recommendations on thresholds.
- As an example, evaluate the Prime Shift data for 2 weeks and set the threshold at 1 standard deviation about "normal". Adaptive monitoring uses Detailed data to evaluate and make recommendations on thresholds. Therefore, you need to keep a reasonable duration of Detailed data in order to perform the evaluation. The duration depends on how the shifts are defined. If you define shifts that include “day of week”, then you need to keep the data longer to get an effective analysis of the data. If you are looking only at "Prime Shift" for all weekdays, then you do not need to keep the data as long.
- Keep 7 to 30 days of detailed data when comparing all work days. If you compare Monday to Monday, then you need to keep the Detailed data much longer to be able to establish a trend. When comparing a specific day of the week, you will probably need to have at least 60 days of data. Before configuring Adaptive Monitoring, consider the use of the data. There is no value in performing Adaptive Monitoring on certain types of data, such as disk space. You must want to set a static threshold on either the % free space or the amount of disk space available. But CPU monitoring is an excellent candidate for Adaptive Monitoring because it can be very beneficial to learn whether a server is behaving abnormally.
- Agent and Attribute Group Considerations
- Each Agent and each Attribute Group must be considered separately when defining Historical Collection. Many Tivoli Monitoring products have defined a set of best practice historical collections. They do not include the summarization and pruning intervals, but are a good place to start when setting up historical collection.
- When looking at these recommendations, consider whether you plan to use adaptive monitoring, short term problem determination, long term reporting, or capacity planning and predictive analysis. This must be taken into account when configuring the summarization and pruning Intervals.
Parent topic:
Summarization and pruning configuration