For our Demand plans, we use m5.xlarge instances through AWS . Each one comes with 4 vCPUs, 16 GB of RAM and supports up to 10 Gbps. To give you an indication, these are our recommended figures for the number of users that can be run on one node for each of the tools we support:
JMeter: 1000 users
Gatling: 1000 users
Selenium: 5 users
Flood Element: 40 users
Since these figures are based on resource utilization, they can vary depending on your script/application. For this reason, we recommend that you use these numbers as a rough estimate to start with, and then proceed to run tests to find your baseline.
When you're hosting your own infrastructure, you can choose a variety of instances depending on the ones you choose to use on your own AWS or Azure account.