Amazon EC2 now has a first-party health signal for the application layer, not just the virtual machine underneath it. AWS announced application status checks for EC2 on August 17, giving site operators a managed way to test whether a workload is actually accepting traffic.

The change matters because the old green light on an instance could still hide a broken hosted service. AWS says the new check can detect cases such as a web server that stops accepting requests, a Docker daemon that is not running, incorrect networking configuration, or a network interface that is no longer passing traffic.

Operators create the check by choosing a protocol, port and path, plus the HTTP response codes that count as healthy. EC2 then sends HTTP or HTTPS requests to the selected endpoint every 60 seconds and reports the result alongside the familiar system and instance status checks.

For hosting teams, the most operationally important piece is the Auto Scaling integration. AWS says Auto Scaling groups can act on application status and replace instances when the application reports unhealthy, moving a common web-hosting runbook closer to native cloud infrastructure.

This does not eliminate the need for external monitoring: customer-visible uptime, DNS, CDN, TLS and regional reachability still need an outside-in view. But it narrows a longstanding gap between infrastructure health and application health inside EC2, especially for fleets that already rely on tags and Auto Scaling groups to manage web tiers.

Source: AWS, “Amazon EC2 introduces application status checks”.