The performance of a BMC ProactiveNet Server depends on several variables, and the performance and scalability guidelines and recommendations address these. Substantial changes in any of the following variables can change the expected scaling and performance numbers:
- Total number of monitor instances
- Total number of attributes.
- To ease the load on the BMC ProactiveNet Servers, ensure that you send to the BMC ProactiveNet Integration Service only the attributes that are absolutely necessary.
- For binary attributes (such as availability of a monitored process) or text messages (such as log messages), send the data as alerts instead of monitored data.
- Service model content, including the following considerations:
- Number of service models
- Number of configuration items (CIs) per service model
- Depth of a service model
- Types of monitored devices
- Number of key performance indicators (KPIs)
- Retention period of data and events
- Number of adapters and monitors on each BMC ProactiveNet Integration Service
- Polling frequency
- Number of concurrent users
- Number of external events per day
- Number of alarms and abnormalities per day
- Disk I/O of the database server
- Network bandwidth utilization
- Hardware configuration, including the number of CPUs, amount of memory, and storage technology (such as SAN or RAID)
Do not use 32-bit hardware to deploy a BMC ProactiveNet Server. No performance or scalability testing is performed on 32-bit hardware. If you are using 32-bit hardware and want to upgrade to version 9.0 of BMC ProactiveNet, upgrade to 64-bit hardware before upgrading the BMC ProactiveNet software.
Adapters for Microsoft System Center Operations Manager (SCOM), BMC Transaction Management Application Response Time (TM ART), and BMC Portal were not tested for scalability in BMC ProactiveNet 9.0.