Improving backup and restore times for large EC2 instances or EBS volumes
Backup and restore performance
You can improve the performance of your EC2 and EBS backup and restores. When Clumio backs up and restores EC2/EBS data, it uses AWS APIs, specifically GetSnapshotBlk and PutSnapshotBlk. These AWS APIs have a default soft limit of 1000 RPS (requests per second).
Clumio recommends improving backup and restore performance by increasing the limits of these APIs. Increasing these limits is particularly helpful if you are protecting either a large number of EC2 instances, or a few large EBS volumes with high rate of data change.
Increasing GetSnapshotBlk and PutSnapshotBlk API limits
An AWS support ticket will need to be raised in the account for which you wish to increase the limits for GetSnapshotBlk and PutSnapshotBlk APIs.
An AWS support ticket format for this is provided below. Text highlighted in RED will need to be updated.
Subject:
Request to raise GetSnapshotBlock and PutSnapshot Block API Limits
Body:
AWS Support - We are currently with Clumio to back up EC2 instances / EBS volumes. I would like to request an increase to the default GetSnapshotBlk and PutSnapshotBlk API limits to 4000 RPS for the following regions:
- {REGION NAME}
- {REGION NAME}
Backup operations will have RPOs ranging from 1 - 24 hours with multiple read/write operations executed in parallel. The daily change rate of EC2 instances / EBS volumes is expected to be between {CHANGE RATE IN GB}. Clumio will spread out API traffic associated with the backup/restore operations throughout a configured backup window.
Questions?
Reach out to [email protected] with any questions.
Updated 7 months ago