Auto Scaling

A scale group is a pool of E2E nodes built from one of your saved images and one plan. You set a minimum and a maximum number of nodes, and E2E adds and removes nodes between them, following an elastic policy, a scheduled policy, or both.

• An elastic policy adds one node when CPU use is above the value you set, and removes one when it is below a second, lower value.
• A scheduled policy uses cron expressions to set the number of nodes at times you choose, one time to scale up and one to scale down.

Every node in the scale group starts from the saved image, so save the image from a node that already runs your application at boot. To give your users one address for all the nodes, put an Application Load Balancer in front of the scale group.

What you can do

• Create a scale group from a saved image and a plan, with its node counts, policy, VPC, encryption, and security group.
• Set an elastic policy that scales on CPU use.
• Set a scheduled policy that scales up and down at set times.
• Resize, stop, or start a scale group, and change its minimum and maximum.
• Put a load balancer in front of a scale group with an Application Load Balancer.
• Manage a scale group’s network and security groups: its VPCs, its public IP, and the firewall rules on its nodes.
• Monitor a scale group: its status, its nodes, their graphs, the audit logs, and the activity timeline.
• Sign in to scale group nodes over SSH or on Access Console.
• Delete a scale group and all its nodes.
• Work with scale groups from a script, with the API operations such as as-list, as-create, and as-resize.

Where to start

To create your first scale group, follow Get started with Auto Scaling. For fixes, see Troubleshooting Auto Scaling; for short answers, see Auto Scaling FAQ, which also covers what a scale group costs.