Monitoring VMs that use recycled IP addresses

When performing auto-discovery of network resources to monitor (hardware devices, virtual machines, cloud resources, etc.), Netreo relies heavily on the IP addresses used by those resources. If, while scanning the network, an unknown IP address is found; Netreo assumes the resource using that IP address is also unknown, and thus new. It then begins the process of adding the resource as a managed device.

To prevent a scenario where Netreo thinks that a newly discovered virtual machine (VM) -- using a recycled IP address from a previous managed device -- is an existing managed device, causing it to concatenate the new device's incoming performance data to the data of the original device, Netreo performs the following deduplication check.

1. Check the UUID of the newly discovered VM against the UUIDs of existing managed devices in Netreo:
   1. If a matching UUID is found, associate the newly discovered VM with that existing managed device and update the VM name and/or IP address as necessary. Monitoring continues as normal and all newly collected performance data is appended to the historical data of the existing device with the matching UUID.
   2. If no matching UUID is found, check for matches to (in order):
      1. MAC address
      2. VM name
      3. IP address
   3. If a match (of any item above) is found:
      1. If the matching device has no UUID associated with it in Netreo, associate the new VM with that existing managed device. Update UUID, VM name and/or IP address as necessary. Monitoring continues as normal and all newly collected performance data is appended to the historical data of the existing device.
      2. If the matching existing managed device does have a non-matching UUID associated with it, device addition for the new VM fails and it is not added to Netreo. The failed device may then be found on the Failed tab of the (Device Management Dashboard) Interrogation page. This event is also recorded in the Netreo audit log.

This deduplication is important to prevent the historical performance data of two unrelated VMs from being combined into a single database, with no way to tell which data belongs to which VM.