Move EBS to Glacier

The EBS snapshot archive is a fully AWS-managed service. If you are only keeping one snapshot of the volume, the cost is about one quarter (1/4) of the cost of a regular snapshot.

Note, however, that the cost advantage only applies when you keep no more than one snapshot of the volume. Normal snapshots made of the same EBS volume are incremental, only keeping one copy of each unique data block. By contrast, snapshots moved to the archive tier are always full snapshots. If you keep multiple snapshots taken of the same volume at different times, and there are minimal numbers of changed data blocks between them, the archive tier can quickly get surprisingly expensive, because every data block that hasn't been changed between the different snapshots will be stored and charged as a separate copy of the data block.

Even when you are keeping only one snapshot, the S3 Deep Archive storage class is much cheaper still. You can attach the EBS volume to a separate Linux instance and use the dd tool to copy its contents to a file on a temporary disk. Once the full hard drive image is stored in the file, you can upload it to S3 in the Standard storage class, just like any regular file, and set a lifecycle rule to transition it to the Deep Archive class one day later, for example. Simple dd parameters usually suffice. You could perform the copy with something like (assuming the volume is /dev/sdc2 in this example, check yours with lsblk):

dd if=/dev/sdc2 of=/mnt/exportdisk/my-image-file-name.dat bs=16M status=progress

If the data is particularly important, you should be very careful to make sure that you know how to restore the disk successfully from the file and have tested restoring it to a completely new EBS volume that didn't previously contain the data. This dd copy process is fully manual and completely up to you to do correctly. If you accidentally copy the wrong volume or only part of the right one, you can easily lose some or all of your data permanently, and even AWS won't have any extra backups to rescue you. With the standard EBS snapshots (whether in the archive tier or not) that are a fully AWS-managed service, you don't need any particular skills beyond choosing the correct volume from which to create the snapshot.