- AWS notifications in Microsoft Teams - Fri, Jun 5 2020
- AWS daily spend CloudWatch notification alert - Mon, Mar 2 2020
- ScriptRunner Version 2019R2: PowerShell scripting for teams with self-service - Wed, Dec 4 2019
How is this changing? ^
Currently you have limits for each EC2 instance by type, limiting you to running a maximum of 20 On‑Demand instances across an instance family. Going forward, Amazon will limit On-Demand EC2 instances by virtual central processing units (vCPUs) only.
This change on how limits occur will now make it easier because you have just one limit to manage all of your EC2 instances. This helps remove the confusion of having limits for the many EC2 instances when it comes to working with such features as autoscaling.
Can you use this new feature before October 24? ^
AWS gives you the option to opt in and use this feature before October 24.
To do so, sign into your AWS console and navigate to the EC2 resource. On the left-side menu, select Limits.
The Limits front screen will prompt a message informing you that the new feature change is coming, and if you would like to use this feature, you can Opt in with the corresponding button.
After opting in, you will be able to use the new features in around 15 minutes.
How AWS calculates the new limits ^
An instance type in AWS describes how much it assigns compute, memory, and storage capabilities to an EC2 Instance. It groups these instance types in instance families based on these different capabilities.
It measures the limits for On-Demand instance vCPUs fixed when running. To break this down, instead of being limited to the number of instance types, you can choose more of one type and less of another. This gives you greater flexibility and freedom to mix and match instance types.
A couple of tools might help:
- Inside the AWS console (under EC2 and limits) is a limits calculator. The tool lets you calculate how many vCPUs you require to launch your On-Demand instances.
Calculate vCPU limits with the simplified vCPU calculator
- Probably the best resource to find the number of vCPUs each instance has is the Amazon EC2 Instance Types link (see below). The site provides comprehensive details on all AWS EC2 instance types.
- Another useful resource is Amazon's FAQs on the new limits. Below is a table from this site that shows the vCPUs for the main common instances used.
The link to find this information is here.
Service quotas ^
Service Quotas allow you to view and manage your quotas from your AWS account. Also, you can track and see the maximum number of quotas (or limits) you can create in your account. You can request an increase in quota values should you need it.
Once the change happens on October 24, there will be only five different On-Demand instance limits. One limit will govern the usage of standard instance families like A, C, D, H, I, M, R, T, and Z. The other limit is per each specialized instance family, which are:
- FPGA (F)
- graphic-intensive (G)
- general purpose GPU (P)
- special memory optimized (X) instances
As stated, you can manage these limits with the Service Quotas console. You will also be able to monitor EC2 usage with Amazon CloudWatch metrics integrations and use the other good stuff CloudWatch provides like alarms.
You have until October 24 to opt in before AWS initiate a hard cutover. Opting in now gives you the control to manage your EC2 instance limits before they enforce the change, so if this is something that might affect you or your company, it is worth thinking about opting in earlier!