Configuring probes on Cisco devices

Probes may be set on the polled Cisco routers though the command line interface. However, certain guidelines apply to their creation in order to ensure polling result correctness.

There are two different ways to poll the IP-SLA probes on the Cisco devices. One can either choose to poll the statistics tables, located in the rttMonStats branch of the Cisco- RttMon-Mib mib or poll the last operation result tables, located in the rttMonHistory and rttMonLatestOper mib branches. Both ways offer pros and cons which are as follows:

	Statistics Tables	Latest Operation Tables (HTTP / Jitter)	History Operation Table (other probes)
Polling period	Long: 5 minutes or more is OK.	Short: Must be as short as the most frequently occurring probe on the polled devices, otherwise there will be missing data points.	Long: 5 minutes or more is OK if multiple history lives and buckets are setup for the probe. This increases memory usage on the polled device. Short: Must be as short as the most frequently occurring probe on the polled devices, otherwise there will be missing data points. This is the case when history buckets and lives are set to 1 to reduce memory usage on the polled device.
Data volume	Low, since the polling period is longer.	High, since the polling period is shorter.	High or Low, depending upon the case.
Collecting host CPU utilization	Low	High	High or Low, depending upon the case.
Polled device CPU utilization	Low, might have a slightly noticeable effect on the device CPU utilization statistics.	High, will have a visible effect on the device CPU utilization statistics.	High or Low, depending upon the case.
Data granularity	Low: Averages are calculated on the polling period. Suitable for performance trending.	High. Suitable for troubleshooting specific problems, occurring at a precise time.	High. Suitable for troubleshooting specific problems, occurring at a precise time.