Configurations Affecting Performance
Connecting to vCenter vs. ESX/ESXi host
Generally connecting ESX/ESXi host directly has better performance. But user must be notified with possible impact on production environment.
Using vCenter is recommended if it is available.
Running Relator in Dynamic Pool vs. Spawning New Monitoring Engine Process
The option ‘Spawning New Monitoring Engine Process‘ gives the best isolation between monitoring functionalities. But it puts the heaviest load on the system, as each new process requires new authentication to the vCenter.
Using Dynamic Pool is recommended.
Size of Shared Monitoring Engine Process Pool
The larger pool size, the more is concurrency on monitoring engine side. Theoretically it should provide better performance. But if vCenter becomes bottleneck, the performance can deteriorate instead.
Number 3-15 is recommended.
- Dynamic Scheduling vs. Fixed Starting Time
Dynamic Scheduling allows spreading the monitoring task among the time line. It averages out the load.
Unless Turbo Mode is turned on, using Dynamic Scheduling is recommended.
- Turbo Mode
This new feature is introduced in 2.0A-1204-B.
In large environment, especially when monitoring large amount of VMs, Argent for VMware or vCenter can be easily overwhelmed.
Turbo Mode is recommended in large environment with many licensed objects.