AnalyticDB for PostgreSQL in Serverless mode automatically backs up your instance data on a rolling 7-day cycle. If data is accidentally deleted or corrupted, you can submit a ticket to restore it to a previous point in time.
Backup is enabled by default for Serverless mode V1.0.0.0 and later. No configuration is required.
How it works
Each backup cycle consists of two types of backups:
-
Base backup (full backup): A complete snapshot of your instance data, taken at the start of each cycle.
-
Log backup (incremental backup): Captures data added or deleted since the last base backup. Log backups run throughout the cycle without compressing data.
Base backups and their associated log backups belong to the same backup group. The start_time of a base backup marks the earliest point in time to which you can restore data within that group.
When instance data is deleted, it is automatically archived and retained according to the backup policy before being permanently removed.
Manage the backup retention period
The default backup cycle is 7 days. To check the current retention period, run the following statement in your instance:
SHOW rds_cn_vacuum_archive_storage_life;To change the retention period, Submit a ticket and specify the number of days you need. Serverless mode manages retention periods centrally, so adjustments require a support request.
View backup records
Connect to your instance and run the following statement in the postgres database:
SELECT * FROM adbpg_tabbak.adbpg_basebackup_task;The adbpg_basebackup_task table contains the following columns:
| Column | Data type | Description |
|---|---|---|
task_id |
text | ID of the backup task |
sched_id |
text | Scheduler ID of the task |
start_time |
timestamptz | Start time of the backup. This is also the point in time to which data can be restored. |
end_time |
timestamptz | End time of the backup |
backupgroup |
int | Backup group ID. A base backup and all subsequent log backups within the same cycle share the same group ID. |
parent_task_id |
text | ID of the parent task |
task_type |
char | Backup type: S = base backup (full backup), I = log backup (incremental backup) |
failed |
boolean | Backup result: f = successful, t = failed |
Restore data
Self-service restoration is not available in Serverless mode. To restore your data, Submit a ticket to contact the technical support personnel.
Billing
Backup storage is billed as part of disk storage — the backup feature itself has no separate charge. Because backups are stored uncompressed, plan for approximately 68% of your data storage as additional backup storage.
Example: An instance with 1 TB of data that receives 100 GB of daily updates will accumulate roughly 700 GB of backup data over a 7-day cycle (100 GB x 7 days).
For disk storage pricing, see Pricing.
What's next
-
Pricing — Understand disk storage fees that apply to backup data