DX - IM - how many profiles snmpget probe can handle - best practices
Any UIM version
Any snmpget probe version
Robot Environment for Running Probes
Robot Machine Considerations
Nimsoft probes are typically run on robots that connect to a hub which connects to the primary hub. For the listed
probes, the following best practice is suggested for selecting machine configuration for the robot that runs them:
Machine type – physical or Virtual(VM)
HA Configuration – Not required
Machine CPU Compute – 2 to 4 core 64bit Pentium or equivalent with speed of 2.0 Ghz with 50% available
on average
Machine system memory – 4GB to 8GB
Machine disk storage – 72GB available
Machine Network bandwidth – standard LAN connectivity bandwidth e.g. 1Gbps. For better results, robot
machine should be in the same subnet as the devices monitored by the probes running on that robot
machine.
Operating Systems – All robot supported OSs e.g. Windows server os– e.g. W2K8R2 or later, Enterprise
Linux v6.1 or later
Network latency and connectivity Considerations
The listed probes are network latency and packet drop sensitive. During testing we considered a latency of less
than 100ms between the Hub and robot. The latency between the probe and the tested application is recorded as
part of the measurements.
Number of devices or profiles per Probe – Key Considerations
Number of devices or profiles monitored by one instance of the probe also affects performance of the probe and
proactive measures below can assist you in making the probe performance optimal for your environments.
Key Configuration Restrictions
SNMP Get
1. SNMP response delays (due to network or device) will reduce the capacity of the probe.
2. Tests ran with 20% of the devices using SNMP v3. Increased SNMP v3 devices will decrease probe
performance.
3. Recommended polling frequency (and tested) is 5 minutes.
SNMP Get - Probe Metrics Count and Runtimes Considerations