Network Discovery timers vs Network Monitoring timers

One of the first steps that we normally perform during POC is timer tuning for ICMP and SNMP for Discovery and Monitoring services.

Normally Discovery should have different timer values than Monitoring because Discovery operates in a “pessimistic” model when IP address

that is being probed by Discovery engine is likely not to be alive or  not to respond to SNMP therefore timeoute values and retry counts has to be very aggressive

for example 100 ms Timeoute with 2 Retries  for ICMP is normally sufficient. SNMP timer for Discovery have typical values of 1000ms and 1 retry.

Aggressive Discovery timers also reduces amount of traffic being generated and make discovery jobs run faster.

Monitoring Service timers are in opposite spectrum,  as Monitoring service operate in “optimistic” mode where it expects for all devices that are enabled for monitoring

to respond and timers has to be tuned to maximum wait time with ICMP timers as high as 300ms and SNMP timers as high as 5000 ms to support bigger/busier devices like Nexus 7018.