CA APM Sizing and Performance
This guide contains recommendations for hardware sizing and configuration. Requirements vary dependent on the size of the environment being monitored, the level of monitoring (number of metrics collected) and the number of extensions and integrations included in the monitoring solution. The example recommendations for various typical installations provide figures for the minimum hardware outlay which ensures consistent good operation.
apmdevops102
This guide contains recommendations for hardware sizing and configuration. Requirements vary dependent on the size of the environment being monitored, the level of monitoring (number of metrics collected) and the number of extensions and integrations included in the monitoring solution. The example recommendations for various typical installations provide figures for the minimum hardware outlay which ensures consistent good operation.
The sizing requirements for components in production environments, handling large numbers of metrics, can be multiple times greater than testing and development installations.
In this guide:
When discussing hardware sizing and performance, we have a specific meaning for the generic term
Enterprise Manager.
We use the term when data is specific to a single Enterprise Manager, and clustering is not relevant to the topic. However sometimes Collectors and MOM Enterprise Managers perform different functions. The functions require different sizing capacity guidelines or result in different performance behaviors. In these cases, the term Collector or MOM is used as appropriate.Optimal Hardware Specifications and Guidelines for an Enterprise Manager
Familiarize yourself with the CA APM product and capacity-planning basics. Then use these summary component and resource recommendations as a starting point. This table describes one recommended Enterprise Manager configuration and is not intended to convey CA APM limitations or requirements.
Hardware Specification | Optimal Guideline |
Number of Enterprise Managers per server | 1 |
Operating System | Red Hat Enterprise Linux 5.x/6.x/7.x |
CPUs | Eight cores |
Physical RAM | 8 GB |
Disk I/O Subsystem | The disk I/O subsystem restrictions apply to all available storage choices. For example, local disks and external storage solutions such as SAN. The OS resides on a separate physical disk from CA APM data. Each SmartStor database resides on a dedicated physical disk. The Enterprise Manager heuristics database (variance.db) and Transaction Event database (traces.db) files can reside on the same physical disk. However, these databases cannot be on the same disk as SmartStor to avoid I/O conflicts with SmartStor. We recommend using SSDs. |
Hardware Recommendations
The following points expand on the hardware recommendations:
- Enterprise Managers need a minimum of four CPU cores to perform key operations. We recommend eight or more CPU cores.
- Enterprise Managers perform many calculations that involve floating-point math. Therefore Enterprise Managers benefit from the x86 chip design for Xeon or Opteron chips more than the RISC chip design. The RISC design is used for SPARC and Power5 chips. For example, a Power5-based server has the same Enterprise Manager CPU and disk requirements as a Xeon- or Opteron-based server. However, the Power5-based server has 20 percent lower capacity.
- Be sure that no other applications or processes run on Enterprise Managers. This requirement helps the Enterprise Manager to avoid conflict with the system resources that Enterprise Managers require.
- Be sure that there is no disk contention for SmartStor.
- To run multiple CA APM instances on the same ESX server, provide each instance with the needed resources. Each instance requires the recommended CPU and disk I/O bandwidth simultaneously during the same 15-second data harvest intervals.