Analyzing the performance and response time of a system or business driver
The following types of time analyses are available in BMC Helix Continuous Optimization:
Performance versus time analysis
A performance versus time analysis (PTA) shows the values of one or more performance metrics over time. It can also show multiple metrics belonging to different systems with different scales. In some cases, it can include derived entities that are created by rolling up metrics to represent resource utilization at a defined abstraction layer.
The main objectives of a performance versus time analysis are:
- Identify the most critical resources to be used for what-if analysis
- Create a performance baseline for the system based on historical data to:
- Verify or change the time frame to be used in further analysis
- Perform trend analysis (linear, moving averages)
- Detect periodical behavior (daily, weekly, monthly)
- Remove outliers
- Determine the positions of recurrent spikes and their values
The following figures are examples of performance versus time analysis plots.
Performance versus time analysis example 1: The behavior of CPU, memory, and disk utilization over time
Performance versus time analysis example 2: Day profile of business driver units consumed
Load versus time analysis
Load versus time analysis (LTA) shows the behavior of one or more business driver metrics with respect to time. An example is the number of transactions over a specific period.
The main objectives of an LTA are:
- Characterize a subset of business drivers by:
- Business value delivered by the application.
- Dependencies with other business drivers.
- Create a time frame to be used in Performance versus time analysis for:
- Trend analysis (for example, linear, moving averages)
- Detecting periodical behavior, intervals, or cycles (for example, daily or weekly night batch activities).
- Determining positions of recurrent spikes (demand peaks or long periods) and their values.
- Create a baseline for characterized business drivers to evaluate the average and variance calculated on historical data in the identified time frame.
The following figure shows an example of different business drivers over a period:
A Load versus time analysis plot
Configuration data analysis
A configuration data analysis shows historical values of system configuration data for the specified period. For each change to configuration data, a column listing all values on that date is displayed, and the changed values are highlighted in bold.
The following figure illustrates the result of a configuration data analysis:
Results of a configuration data analysis