shiichan

AWS Elastic Disaster Recovery gets Recovery Plans — say goodbye to manually managing server launch order!

Hi, I'm Shii-chan! Today I want to share a nice update to AWS's disaster recovery service.

AWS What's New aws.amazon.com

What was announced?

According to AWS What's New, AWS Elastic Disaster Recovery (AWS DRS) now includes a new capability called Recovery Plans. It automates the sequential launch of multi-server applications made up of databases, application tiers, and supporting services during recovery and drills.

The story so far

Previously, when recovering an application made up of multiple servers, you had to launch servers one at a time and track the dependencies between them yourself. If you brought up the application tier before the database was ready, things could break. And the tricky part is that this work often had to happen during an actual, high-stress disaster scenario, which made it easy to make mistakes in the launch order.

What changes

With Recovery Plans, you define the recovery sequence once and then run it with a single action whenever you need it. You group servers into sequential steps, with configurable wait times between them, so your application recovers predictably and consistently every time. This reduces recovery time and eliminates common errors that occur when coordinating server launch order under pressure.

Dive Deep

Here's what the announcement specifically calls out:

  • Group servers into sequential steps with configurable wait times between them
  • Run plans in a non-disruptive drill mode to validate your recovery procedures without an actual failover
  • Add approval steps where you want a person in the loop before proceeding
  • Monitor recovery progress in real time
  • Available in all AWS Regions where AWS DRS is offered
  • No additional cost beyond standard DRS usage

For setup details, AWS points to the AWS Elastic Disaster Recovery User Guide.

Wrap-up

  • AWS DRS now offers Recovery Plans to automate the launch order for multi-server applications during recovery
  • Previously, servers had to be launched one at a time with dependencies tracked manually
  • You group servers into steps with configurable wait times, define the plan once, and run it with a single action
  • Supports non-disruptive drill mode, approval steps, and real-time progress monitoring
  • Available in all Regions at no additional cost

This is especially useful for infrastructure and SRE folks who protect multi-server systems with AWS DRS and want to standardize and automate their disaster recovery procedures.