EC2 On-Demand Instance Limits

VCPUs (virtual central processing units) are assigned for running the on-demand instances in your AWS account. When you run these instances, there is a limit set for them. The limit is measured on the basis of the number of these vCPUs used. That is called the vCPU-based limits.

You can run up to five vCPU-based instance limits on Amazon EC2 with each of the limit having a different capacity depending on the instance family.

Yes, all the EC2 On-Demand Instance Limits are set as per the regions. They are set on the basis of per region.

Yes, the limits can change over the time. Amazon monitors your usage for the EC2 instances and based on the usage per region, the limits are automatically increased.

Generally, Amazon itself increases the limits based on your use as per its monitoring. However, if you still wish the limits to be increased further, you can request for the same by filling the increase limits form on Increase Limits Page in Amazon EC2 Console.

There is a simplified vCPU calculator available for calculating and computing the vCPU limits for your entire AWS account. Alternatively, you can calculate the limits through vCPU mapping for each EC2 Instance Types.

No, vCPU limits do not apply when you are purchasing or requesting reserved instances or spot instances. These limits only apply to the EC2 On-Demand Instances.

You can view the current limits on the EC2 Service Limits Page in Amazon EC2 Console, or in the Service Quota Console and APIs.

There shall be no effect on the running instances due to opting for the vCPU-based instances. The running instances will function as normally as they did before.

No, you can still launch the same number of instances as you could launch with the count-based instance limits when you are using the vCPU-based instance limits.

There are several ways to track and view your usage against the limits. You can use the Trusted Advisor and Limit Monitor for tracking the usage. Alternatively, you can integrate with the Amazon CloudWatch metrics and view the limit usage in the Service Quotas Console which also allow to configure alarms that warn about the approaching limits.

No, with the vCPU-based limits, the total instance limits do not govern the usage. So, DescribeAccountAttributes would not run the max-instance values and will not be supportable. You can instead use Service Quotas APIs to retrieve the information for your EC2 limits.

EC2 use is calculated based on the per hour or per second usage, the AMIs you are running, and the launched instance size and type. So, your monthly bill will see no changes as such.

Generally, the vCPU limits are available in all the commercial AWS regions. However, AWS China Regions (Beijing and Ningxia) are an exception. For these regions, they are available in AWS GovCloud (US).