Schedule a controlled stop - Best practices for assessing applications to be retired during a migration to the AWS Cloud

Schedule a controlled stop

In your migration plan, be sure to schedule time for a controlled stop during the migration process. A controlled stop pauses the migration process to identify the potential for disruption if an application is retired. It simulates the application’s retirement, and allows you to observe the consequences. When the controlled stop period is completed, the migration can easily resume.

The controlled stop approach varies depending on the type of application and the associated processes you are working with. Common controlled stop patterns include:

  • Implementing a host-based firewall to block all traffic, which simulates retirement

  • Pausing a virtual machine

  • Stopping a service on the host

  • Blocking all traffic by using an external firewall

The migration project and application owners need to define the duration of a controlled stop, depending on the application type. For example, if you’re retiring a batch-based workload that only runs once a month or once a quarter, performing a one-week controlled stop might not be enough to determine the impact on other systems.

Continuing with the example from the previous section, another application was connecting to the server that was scheduled for retirement. An initial assessment concluded that there should not be an impact on the upstream servers. A controlled stop can now be carried out to understand the impact.

This controlled stop would be carried out by implementing a host-based firewall to block all traffic, simulating the effect of shutting down the server. If this causes service issues for applications that are scheduled to be migrated to the AWS Cloud, a firewall rule is added and all traffic resumes. After the controlled stop, the server’s retirement is reconsidered due to this service degradation or interruption.