I am confused as to the durability of the EBS volumes on AWS.
As an example, volumes that operate with 20 GB or less of modified data since their most recent Amazon EBS snapshot can expect an annual failure rate (AFR) of between 0.1% – 0.5%, where failure refers to a complete loss of the volume. This compares with commodity hard disks that will typically fail with an AFR of around 4%, making EBS volumes 10 times more reliable than typical commodity disk drives.
Because Amazon EBS servers are replicated within a single Availability Zone, mirroring data across multiple Amazon EBS volumes in the same Availability Zone will not significantly improve volume durability.
Does this mean that the 1/200-1/1000 probability of failure is before the automatic replication or after?
The quoted AFR refers to the end-to-end experience of using an EBS volume. EBS replication is completely transparent to you, and you won't even know when one of the replicas fails and is automatically replaced.