Determine correct sizing for load test VMs (or scale to multiple VMs)

[theotherian]

Currently the load test VM seems to topple over once a reasonable amount of load is placed in the ilpv4.IlpOverHttpOneConnectorSimulation test (15 minutes at 50 users).

At this point, it's unclear if:

• The problem is with the load test hosts
• It's a CPU bound problem (we have 2 vcpus)
• It's a memory bound problem (we're using 6gb and the vm has 7.5gb)
• It's a network bound problem

First steps are to replicate behavior locally and then try expanding the size of the VM until we get a successful result.

If we cannot get a successful result, we need to look at Gatling's ability to spread a test across multiple VMs and launch a cluster via the Cloud Function.