Currently the probe cluster will spread the discovery range IPs over the available probe servers, but if a lot of the probe ranges dont have active devices, the IP addresses are still assigned to a probe meaning anything other then IP discovery will not be balanced.
I would like to see the IP addressed balanced based on the outcome of the IP scans so only active IP addresses are considered.
We only do agentless discovery. The probe clustering in its current form is useless. A probe could have 10,000 IP addresses, but only 500 Windows nodes it's discovering by shell. Another probe could have 1,000 IP addresses, and 1,000 Windows nodes it's discovering by shell. If they were clustered, the busy probe would be given 8,000 IP addresses and potentially up to 500 more Windows nodes. This is not intelligent. Intelligent clustering would balance the workload, so that each probe has 750 Windows nodes AND 8,000 IP addresses. Better yet, the app would analyze other factors and create a highly optimized balance between the probes.
Nice to see the idea. We also have similar idea in internal innocation session to have loadbalance in between Probes and Agents(Smart agent) and have agent to find probe intead of you pre-config the iprange on the probe. The limitation is it must be smart agent... and I am not sure how many customer like to install agent agnist agentless.