How to Adjust AWS ALB Deregistration Delay Beyond 1 Hour

Select Language:

If you need to keep instances running for a longer period—such as 24 hours—for draining traffic before shutting them down, you’ll find that the load balancer’s maximum deregistration delay of 3600 seconds (or one hour) can be a challenge. Since this limit cannot be increased through official settings, you’ll need to use alternative methods to achieve a longer drain time.

One effective approach is to manage the deregistration process manually. First, you should remove the instance from the target group so it stops receiving new traffic. Then, keep a close eye on the current connections on that instance. The goal is to allow existing connections to finish naturally without new ones starting. Once all active connections have completed—after your desired amount of time—you can proceed to terminate the instance.

Another helpful method is gradually reducing the traffic. This can be done by setting up weighted target groups or using DNS-based traffic steering. Instead of removing an instance all at once, you slowly shift traffic away from it over time, giving users a smoother experience and more control over the shutdown process.

You can also improve this process by incorporating connection management within your application. Within your code, you can implement logic to refuse new connections while allowing ongoing ones to finish. This approach ensures your application handles shutdowns gracefully, regardless of the load balancer’s default delay settings.

Lastly, consider combining multiple layers of connection management. For example, use the load balancer’s connection draining alongside auto-scaling procedures or deployment scripts that manage connections at the application level. This multi-layered approach can help extend the effective drain period beyond what the load balancer alone allows.

Since the 3600-second limit is a fixed service constraint, overcoming it requires planning and custom workflows. By manually controlling connections, gradually shifting traffic, and managing application behavior, you can effectively implement longer drain periods without relying solely on the load balancer’s built-in features.

Sources and additional information can be found in the official AWS documentation on modifying target group attributes and configuring connection draining.