I have create a AWS EC2 instance to run a computation routine that works for most cases, however every now and then I get an user that needs to run a computation routine that crashes my program due to lack of RAM.

Is it possible to scale the EC2 instance's RAM and or vCPUs if required or if certain threshold (say when 80% of RAM is used) is reached. What I'm trying to avoid is keeping and unnecessary large instance and only scale resources when needed.

CodePudding user response：

It is not possible to adjust the amount of vCPUs or RAM on an Amazon EC2 instance.

Instead, you must:

• Stop the instance
• Change the Instance Type
• Start the instance

The virtual machine will be provisioned on a different 'host' computer that has the correct resources matched to the Instance Type.

A common approach is to scale the Quantity of instances to handle the workload. This is known as horizontal scaling and works well where work can be distributed amongst multiple computers rather than making a single computer 'bigger' (which is 'Vertical Scaling').

The only exception to the above is when using Burstable performance instances - Amazon Elastic Compute Cloud, which are capable of providing high amounts of CPU but only for limited periods. This is great when you have bursty needs (eg hourly processing or spiky workloads) but should not be used when there is a need for consistent high workloads.