This article explains how Flexera Elastigroup handles locked instances during autohealing. This helps you understand which protections apply and when Elastigroup may still replace a locked instance.

How instance locking works

You can lock instances in Elastigroup to stop them from being scaled down during autoscaling. Locking protects the instance for a set period and prevents scale-down actions.

Locking does not protect an instance from:

• A Spot interruption

• Node replacement during a cluster roll

How autohealing behaves with locked instances

Autohealing replaces instances that fail the health checks you configure. When an instance becomes unhealthy:

• The unhealthy instance handler starts autohealing

• Autohealing attempts to detach the instance

• Detaching an instance does not consider instance protection

Because of this workflow:

• Autohealing does act on locked instances

• Locked instances can still be detached and replaced when unhealthy

• Instance protection applies only to scale‑down, not to autohealing

Why protection doesn’t apply

Instance protection focuses on preventing autoscaling scale‑down actions. Autohealing uses a separate mechanism designed to ensure your group stays healthy. That mechanism ignores instance protection when an instance is unhealthy.