The most common questions asked are:

• How much resources do I need to enable Network Observability?
• Given the size of my cluster, do I have enough resources to enable Network Observability?

This Jira issue is to test and come up with answers to these questions.  The CPU resource is the most critical and tied to memory since selecting the number of CPUs for a node also determines the amount of memory (although there are varying options).

The tasks are:

1. Determine the peak CPU usage.  Then see what percentage of peak usage is needed by Network Observability.  For example, if you have 40 nodes with 16 vCPUs per node and peak usage is 80%, that is 512 vCPUs.  If Network Observability needs max 50 vCPUs, then that is slightly less than 10% of peak CPU.

2. Do similar analysis for memory.

3. Note how much storage is used on a daily basis.

4. Test with different testbeds.  See NETOBSERV-224, NETOBSERV-483, and NETOBESRV-484.