AWS Aurora Snapshot and AWS Backup Service

It appears you have some excellent questions about AWS backups, particularly concerning Aurora and RDS snapshots. Let's clarify some of the points you've raised.

Your understanding of AWS snapshots as full backups is generally correct. When AWS takes a snapshot, whether it's for Aurora or other RDS databases, it's a point-in-time full backup of the data. For an Aurora cluster, the snapshot captures the entire cluster volume, encompassing all data across its instances. When you restore from an Aurora cluster snapshot, you restore the entire cluster. For a standalone PostgreSQL instance (non-Aurora), the snapshot is of that specific instance, and a restore will create a new instance based on that snapshot. You are also correct that you typically don't see "non-Aurora PostgreSQL Clusters" in the same way you see Aurora clusters; traditional RDS PostgreSQL operates with individual instances, although you can set up high availability with Multi-AZ deployments.

Regarding the different categories of snapshots you see in Aurora and RDS > Databases - Snapshots:

• Manual: These are indeed snapshots you initiate manually. You create them explicitly when you want a specific point-in-time backup.
• System: These refer to automated backups performed by AWS. When you enable automated backups for an RDS instance or Aurora cluster, AWS regularly takes snapshots and retains them for a specified backup retention period. These are distinct from the "Automated backups" section you mentioned, which often provides a chronological view of these system-generated backups, allowing for point-in-time recovery.
• Shared with me: This category is for snapshots that have been shared with your AWS account by another AWS account. Your observation that it's currently empty is typical if no one has shared snapshots with you.
• Public: As you noted from the documentation, these are snapshots that have been made publicly available by other AWS accounts. You won't be billed for these, and they are generally used for sharing specific database configurations or data for public consumption, though caution should be exercised when using public snapshots due to potential security implications.
• Exports in Amazon S3: This section shows snapshots that have been exported to Amazon S3. AWS allows you to export snapshots to S3 for long-term retention, analytics, or disaster recovery purposes. If you haven't configured any exports, this section would be empty.

The "Backup service" category is indeed a source of confusion for many.

The Backup service refers to backups managed by AWS Backup. AWS Backup is a centralized, managed service that makes it easier to back up your data across AWS services (including RDS, Aurora, EBS volumes, EC2 instances, S3, etc.) and on-premises using a single console.

Here's a clearer explanation:

• Is this something that we/customer will have configured ourselves or is it an AWS default of some sort? AWS Backup is something you, as the customer, configure. It's not an AWS default for RDS/Aurora backups. While RDS/Aurora has its own built-in automated backup mechanisms (which show up as "System" snapshots), AWS Backup provides a more unified and policy-driven approach to managing backups across multiple services. You would define backup plans in AWS Backup, specifying things like backup frequency, retention periods, and target backup vaults.
• There is a DB instance or cluster associated with the snapshot name in the Backup service. Are these backups of the snapshots themselves? How are they used? Are they used behind the scene when a restore is initiated? No, these are not backups of the snapshots themselves. When you configure AWS Backup to back up your RDS instances or Aurora clusters, AWS Backup creates and manages these snapshots as part of its backup plan. These are essentially another set of full backups of your database, but they are managed by the AWS Backup service according to the rules you've set up in your backup plans. When you initiate a restore from a backup listed under "Backup service," AWS Backup uses that specific backup to restore your database. They are not simply "used behind the scenes when a restore is initiated" if the restore is from a "System" or "Manual" snapshot; rather, they are the primary source for restores initiated through the AWS Backup console or API.

In essence, you have two primary ways to manage backups for Aurora and RDS:

1. Native RDS/Aurora automated backups and manual snapshots: These are managed directly within the RDS console and provide the "System" and "Manual" categories.
2. AWS Backup: This is a separate, centralized service that can take backups of your RDS/Aurora instances (and many other AWS resources) and manages them under the "Backup service" category in the snapshots list. It offers a more comprehensive and auditable backup solution for organizations with diverse backup needs.

You can find more detailed information on AWS Backup and its integration with RDS/Aurora in the official AWS documentation:

• What is AWS Backup?
• Working with Amazon RDS in AWS Backup

Hopefully, this provides a clearer explanation of the different backup types and the role of AWS Backup!

The "Backup service" tab in the Aurora/RDS Snapshots section refers to backups that are managed by AWS Backup, which is a separate AWS service that provides centralized and automated backup management across multiple AWS services.

Let me clarify the different types of backups for Aurora and RDS:

1. Manual snapshots: These are user-initiated backups that are retained until you manually delete them.

2. System snapshots: These are automated backups taken by RDS/Aurora during your specified backup window. They're retained for your configured retention period (1-35 days).

3. Backup service: These are snapshots created and managed by AWS Backup. Although they appear similar to manual snapshots, they have some key differences:

• They have names with the format "awsbackup:backup-job-number" for RDS or "awsbackup:job-AWS-Backup-job-number" for Aurora
• They don't count toward your DB snapshot quota
• They're created through AWS Backup plans that you or someone in your organization has configured

AWS Backup is not enabled by default - it's a service that needs to be configured separately. It allows you to create backup plans with specific schedules, retention policies, and lifecycle rules that can back up multiple AWS resources (not just databases) in a centralized way.

For Aurora specifically, AWS Backup can also manage automated (continuous) backups. These have names with "continuous:cluster-AWS-Backup-job-number" and can be used for point-in-time recovery.

Regarding your understanding of cluster vs. instance backups:

• For Aurora: Backups are at the cluster level and include all instances in the cluster
• For standard RDS: Backups are at the instance level since there's no concept of a cluster (except for Multi-AZ DB clusters)

The "Exports in Amazon S3" tab shows snapshots that have been exported to S3, which allows you to access the data directly outside of RDS.
Sources
Overview of backing up and restoring an Aurora DB cluster - Amazon Aurora
Using AWS Backup to manage automated backups for Amazon RDS - Amazon Relational Database Service
Amazon Relational Database Service backups - AWS Backup
RDS Snapshot vs Automated Backups | AWS re:Post