Auto Scaling Documentation

Use these guides to plan, create, manage, and troubleshoot Auto Scaling scale groups in MyAccount. Most workflows start from Compute > Auto Scaling in the left navigation.

Auto Scaling grows and shrinks a pool of compute nodes automatically against a policy you define, so your application stays responsive during peak demand and costs less during quiet periods. The Scaler launches and terminates nodes from a saved image, registers and de-registers them on your load balancer, and scales on CPU, a custom metric, a schedule, or a combination. If you are new to it, start with Auto Scaling Concepts.

Plan and Launch

Auto Scaling Concepts

Scaler, scale groups, group nodes, saved images, plans, policies, watch periods, and cooldown.

Create a Scale Group

Pick a saved image and plan, set node counts, and configure elastic and scheduled policies.

Operate

Manage Scale Groups

Hub for the Details, Nodes, Network, Monitoring, Security, Logs tabs and lifecycle actions.

Scale Group Actions

Resize the active node count, stop, start, and delete a scale group.

Details and Editing

Review the configuration and edit node counts and policies inline.

Node Details

Inspect active and terminated group nodes and open a console.

Network and Load Balancer

Manage the VPC, public IP, and the load balancers fronting the group.

Monitoring

Read CPU, memory, and disk graphs to confirm scaling behavior.

Security Group

Attach, review, and detach security groups for group nodes.

Logs and Timeline

Read the Scaler's state audit logs and the management activity timeline.

Capabilities

Custom Scaling Policies

Scale on memory, network, disk, or any custom metric you publish to the nodes.

Auto Scale Encryption

Encrypt group nodes at rest with LUKS, enabled at creation time.

Autoscale for E1 Series

Storage, IOPS, and hourly billing behavior for E1-series scale groups.

Fix Issues

Troubleshoot Auto Scaling

Fix groups that won't scale, stuck statuses, disabled actions, and custom-metric issues.