To protect against fraud all new AWS accounts have low account limits for launching EC2 instances. This means you will be restricted to launching the smaller instance types (the exact number is account dependant). You can request an increase to the limits by using the Service Quotas tool. This is done per-region so needs to be repeated for each region you want to run in.
To start, search for ‘Service Quotas’ in the top search bar and you should see a similar page as below.
Next, click Elastic Compute Cloud (EC2) - (this is where you can increase other service limits too)
You can now search for ‘on-demand’, which should bring up the following page:
You can now click the radio button next to the ‘Standard Instances’ and click request increase. The request is in vCPU’s, so you should consider that each c5n.18xlarge node has 72 vCPUs. For some of the examples in this workshop, you will need a 240 vCPUs limit, but request to whatever number you need. You will then get an automated e-mail from the AWS support team and they either notify you of it being approved or request further information, depending on the number chosen.
Please repeat this for the G instances (GPUs), which we’ll use to create a visualization node. We would recommend a minimum of 32 vCPU’s, but you may want higher for larger nodes.