Distributed cluster sizing

Distributed configurations consisting of one primary reporting server and multiple remote collectors allow Plixer Scrutinizer to scale beyond the single-appliance ceiling of 500 exporters with a total flow rate of 150,000 flows/s.

This section contains resource allocation guidelines and recommendations for individual appliances in a distributed cluster.

Remote collectors

In a distributed environment, resource allocation for each remote collector should follow the same guidelines/recommendations as that of a single Plixer Scrutinizer appliance:

  1. Use the expected flow rate and exporter count for the collector to determine recommended CPU and RAM allocations for core functions.

  2. Calculate the total additional CPU cores and RAM required to support the features that will be enabled for the collector and exporters associated with it.

  3. Provision the collector with the minimum 100 GB of disk space and the total CPU and RAM obtained from the first two steps.

After the collector has been registered as part of the cluster and is receiving flows, continue to monitor resource utilization via Admin > Resources > System Performance page and make adjustments when necessary.

Primary reporter

CPU and RAM requirements for the primary reporter in a distributed environment are primarily based on the number of remote collectors in the cluster:

Minimum

Recommended

CPU cores

2x the number of remote collectors

4x the number of remote collectors

RAM

2 GB for every remote collector

4 GB for every remote collector

Note

  • The CPU core and RAM allocations above are exclusive of the base resource requirements for the virtual appliance.

  • Depending on the scale of the network, the primary reporter may be subject to additional load due to the volume of Alarms/events being forwarded by the collectors.