If graphs or reports are truncated and not displaying properly, or users are experiencing performance issues at the workstation while graphs and reports are being generated, you may be exceeding non-modifiable graph and report limits. Graphs and reports containing too many data points can adversely impact Enterprise Manager performance.
Generating CA APM reports is expensive in terms of CPU and disk access. The cost is primarily based on two factors:
- Number of graphs (total amount of data)
- Report time period (historical range)
Reports temporarily reduce Enterprise Manager capacity. While reports are running, Introscope users can experience the symptoms of Enterprise Manager overload, such as gaps in workstation graphs, reduced responsiveness, and combined time slices.
- Do not schedule reports containing more than 50 graphs or for historical periods longer than 24 hours during SmartStor reperiodization (occurring by default between midnight and 3:00 A.M.). During reperiodization, Enterprise Manager CPU and disk activity is increased.
- The following limits apply to tables in reports:
- The maximum number of metrics that can be reported in one table is 6000.
- The limit on data size for tables in reports is 6000. This is the limit on the number of data points fetched for a table. The limitation is in place because workstations can issue requests for arbitrarily large amounts of data. Data points are not sent in their entirety but as summary values. Resources are consumed by the metric data service.
- The following limits apply to graphs in reports:
- The default limit for the metric size of graphs in reports is 500. This is the number of metrics allowed to display in a single chart. The limit is more restrictive than tables because charts contain the full time series of data, whereas tables only keep a running total. Also, the number of plots which can be reasonably viewed in a chart is limited.
- The default value for data size limits in graphs in reports is 500. This is the limit on the number of data points fetched for a graph. There is a cost associated with data in graphs in terms of the amount of memory required and performance overhead consumed by creating time series objects. For this reason, the limitation is more restrictive than that of tables. Generally, there is no reason to increase resolution beyond about 400 data points since this will exceed maximum granularity.