Today, we are making available a new option to our “Backup Multiple EC2 Instances” action: Rolling Instance Stops.
Skeddly’s “Backup Multiple EC2 Instances” action is a powerful tool for creating EBS snapshots of your EC2 instances. It can discover your instances automatically, create the snapshots, tag them, and copy them to another region.
As part of this action, it can stop your EC2 instances for the snapshots and start them up again. Until now, all the instances were stopped simultaneously. However, this may have impeded the ability to maintain high-availability of your applications.
Rolling instance stops will group instances based on selected criteria and wait for each “group” of instances to resume running before stopping the next “group”. If you are using multiple actions in order to “time” the stopping of your EC2 instances, then you can combine them into a single action.
There are a few options for rolling instance stops:
The EC2 tag option can be quite powerful and allows you to customize how your EC2 instances are grouped and the order in which they will be stopped.
For example, you can have a tag on your EC2 instances called “Skeddly Backup Order”. Some may have a tag value of “A”, others “B”, others “C”. The actual tag value does not matter. All the instances found by the action will be grouped together that have the same value (for example, all instances with tag value “A” will form a group). Then the groups will be stopped and restarted based on the alphabetical ordering of those values. So those with value “A” will stop and restart, then those with value “B”, then those with “C”.
If the one-at-a-time and availability zone options aren’t good enough, then the EC2 tag option should give you the control you require.
This new ability is available today.