Analysis Reports

When you request an Analysis Report, the Performance Manager analyzes data for the requested time period and nodes against the rules in the knowledge base.

The Analysis Report consists of conclusions for each node and includes
cluster-wide conclusions for a cluster system. Each conclusion is caused by a rule firing. When a rule fires, the Performance Manager reports the problem condition and makes a recommendation for improving performance. All rules are identified with a unique five-character alphanumeric code, such as {M0010}. A rule identifier appears with each conclusion.

In addition to the conclusions, you can request that the Performance Manager list the rule conditions that satisfied the rule firing. The Performance Manager also provides supporting evidence. By default, the Performance Manager provides conditions and evidence when you specify an output file or when you run the analysis process in batch mode. To suppress the conditions and evidence in the report, use the /NOEXPLAIN qualifier.

The conclusions in the Analysis Report for a cluster system are listed in the following order by node:

Each node's local analysis and conclusions, which may include the following conclusions:
- Memory-related
I/O-related
- CPU-related
- Miscellaneous
- Auxiliary
Cluster-wide analysis and conclusions, which may include the following conclusions:
- HSC limitation
- Disk-related
- Lock-related
- Auxiliary

If you request a Performance Evaluation Report along with an Analysis Report, the performance data for each node follows that node's local analysis. The performance data for the cluster appears after the cluster-wide analysis.

Factory analysis reports can be produced only from data collected by the primary data collector, namely that which is associated with the CPD definition.