Skip to main content

Database Backups

Automated backups with recovery options for PostgreSQL and MariaDB databases.

Database Backups

PostgreSQL Backups​

Backup Features​

SailOps provides comprehensive backup protection for PostgreSQL:

  • Automated daily backups - No manual intervention required
  • On-demand backups - Create a restore point before a risky operation
  • Managed retention - Completed backups are retained for up to 30 days
  • Encrypted UI downloads - AES-256 ZIP encryption with a one-time password

How Backups Work​

Daily Base Backups​

  • A physical CloudNativePG base backup is taken daily
  • Retained per your retention policy
  • Restores use a specific completed backup as their restore point

Continuous WAL archiving is not enabled, so arbitrary timestamp recovery between backups is not available.

Retention Policies​

Backups are automatically retained for up to 30 days:

RetentionAvailable restore points
Up to 30 daysCompleted backups retained during that period

Viewing Backups​

  1. Navigate to your database
  2. Click the Backups tab
  3. View available backups and restore points

Backup Information​

FieldDescription
DateWhen the backup was created
SizeCompressed backup size
TypeScheduled, Manual, or Automatic
StatusPending, Running, Completed, Failed

Restoring from Backup​

PostgreSQL restore always creates a new database cluster — your existing database is never modified.

Encrypted Downloads

Completed backups can be downloaded as AES-256 encrypted ZIP files. An encryption password is generated and shown once — save it immediately as it cannot be retrieved later.

Backup Restore​

Restore a specific backup to a new database cluster:

  1. Go to Backups tab
  2. Find the backup you want to restore
  3. Click Restore
  4. Enter a new cluster name using lowercase letters, numbers, and hyphens
  5. Click Restore

The selected completed backup is resolved to its exact Barman recovery ID; the service does not silently substitute the latest backup. The restored database appears immediately as a provisioning database in your project, and the durable operation resumes automatically if UserAPI restarts.

Roles and extensions are part of PostgreSQL physical recovery and cannot be excluded. SailOps copies the current managed-role credentials into target-owned Secrets, then CloudNativePG reconciles those credentials on the restored cluster. Public access and connection pooling are not copied; configure them explicitly after the target is ready. The source database cannot be deleted while an accepted restore still depends on its backup data.

Use Cases
  • Recover from accidental data deletion
  • Restore a backup created before a problematic deployment
  • Create a snapshot for testing

Manual Backups​

Create an on-demand backup from the UI:

  1. Go to Backups tab
  2. Click Create Backup
  3. Wait for completion

Downloading Backups​

Download completed backups directly from the UI:

  1. Go to Backups tab
  2. Find a completed backup
  3. Click the download icon
  4. The backup downloads as an AES-256 encrypted ZIP file (.enc.zip)
  5. An encryption password is displayed in a modal — save it immediately
  6. The password cannot be retrieved later and the modal auto-dismisses after 5 minutes

To decrypt the downloaded backup:

  • Use any ZIP tool that supports AES-256 encryption
  • Enter the provided password when prompted
PostgreSQL download format

The downloaded file contains a physical CloudNativePG/barman base-backup archive, not a SQL dump. After decryption, its entry is named backup-<id>-data.tar with the source compression suffix when present (for example .gz), so its physical format is explicit. It does not include the WAL segments and metadata needed for a standalone restore and cannot be imported with psql. Use pg_dump yourself when you need a portable logical export.

Public API

Backups can also be managed via the Public API (list, create, get, download). The Public API download returns the raw physical archive through a presigned URL; it is not the AES-encrypted ZIP produced by the Dashboard download.

MariaDB Backups​

Automated Backups​

MariaDB databases are backed up automatically on a configurable cron schedule. The default schedule is 0 2 * * * (daily at 2 AM). Backups are stored on S3-compatible backup storage.

Retention​

MariaDB backups are retained for up to 30 days.

On-Demand Backups​

Create a manual backup at any time:

  1. Navigate to your MariaDB database
  2. Click the Backups tab
  3. Click Create Backup
  4. Wait for completion

Restoring from Backup​

Destructive Operation

MariaDB restore is an in-place restore — it overwrites the current data in your existing database. This is different from PostgreSQL, which restores to a new cluster.

Encrypted Downloads

Completed backups can be downloaded as AES-256 encrypted ZIP files. An encryption password is generated and shown once — save it immediately as it cannot be retrieved later.

To restore a MariaDB backup:

  1. Go to Backups tab
  2. Find the backup you want to restore
  3. Click Restore
  4. Type the database name to confirm the operation
  5. Click Confirm Restore

Downloading Backups​

Download completed MariaDB backups directly from the UI:

  1. Go to Backups tab
  2. Find a completed backup
  3. Click the download icon
  4. The backup downloads as an AES-256 encrypted ZIP file (.enc.zip)
  5. An encryption password is displayed in a modal — save it immediately
  6. The password cannot be retrieved later and the modal auto-dismisses after 5 minutes

To decrypt the downloaded backup:

  • Use any ZIP tool that supports AES-256 encryption
  • Enter the provided password when prompted
Public API

Backups can also be managed via the Public API (list, create, get, download).

Backup Best Practices​

Test Your Backups​

Regularly verify backups work:

  1. Perform a test restore monthly
  2. Verify data integrity
  3. Document the process
  4. For PostgreSQL, delete test database clusters when done

Know Your RPO/RTO​

MetricDescriptionSailOps Default
RPO (Recovery Point Objective)Max data lossTime since the selected completed backup
RTO (Recovery Time Objective)Time to recoverDepends on database size and cluster provisioning

Additional Protection​

For critical data, consider:

  • Portable logical exports with pg_dump for PostgreSQL
  • Application-level backups for business-critical data

Disaster Recovery​

Scenarios Covered​

PostgreSQL​

ScenarioRecovery MethodTime
Accidental deletionRestore a completed backup to a new clusterDepends on backup size
Data corruptionRestore an earlier completed backup to a new clusterDepends on backup size
Instance failureAuto-failover (HA)< 30 sec
Zone failureRestore from a completed backupDepends on backup size and provisioning

MariaDB​

ScenarioRecovery MethodTime
Data lossRestore from backup (in-place)5-15 min
Instance failureAuto-failover (Galera)< 30 sec
Zone failureRestore from backup (in-place)10-30 min

Recovery Process (PostgreSQL)​

  1. Identify the issue — When did it occur?
  2. Choose a completed backup — Select the latest suitable restore point before the incident
  3. Restore to new database — A new cluster is created, production is untouched
  4. Verify data — Check data integrity on the new cluster
  5. Switch over — Update application to use restored database
  6. Clean up — Remove old database if appropriate

Recovery Process (MariaDB)​

  1. Identify the issue — When did it occur?
  2. Select the backup — Choose the most recent backup before the incident
  3. Confirm the restore — Type the database name to confirm (in-place, destructive)
  4. Verify data — Check data integrity after restore

Backup Monitoring​

Backup Status​

Check backup health in your database dashboard:

  • Last successful backup - Should be within 24 hours
  • Backup size trend - Unusual changes may indicate issues
  • Backup list - View all available backups

Compliance​

Backup Encryption​

Dashboard downloads are wrapped in an AES-256-encrypted ZIP and transferred over HTTPS. The raw Public API download is not wrapped in this encryption layer.

FAQ​

How long do restores take?​

Typically 5-30 minutes depending on database size.

Can I restore to a different project?​

Not directly. Restore to the same project, then migrate data.

Are backups cross-region?​

The storage location depends on the backup provider configured for the managed database. Do not assume that one backup is replicated across regions.

Can I download my backups?​

Yes, you can download completed backups from the Backups tab. Downloaded backups are AES-256 encrypted ZIP files with a server-generated password shown once during download.

What happens if I delete my database?​

Do not rely on managed backups remaining available after deleting the database cluster. Download or create a portable logical export before deletion when you need independent retention.

Troubleshooting​

Backup Failed​

  • Check database health
  • Verify storage space
  • Review database logs
  • Contact support if persists

Restore Taking Too Long​

  • Large databases take longer
  • Check network conditions
  • Contact support for status

Data Missing After Restore​

  • Verify that you selected the intended completed backup
  • Compare the incident time with the backup creation time
  • Try an earlier completed backup if the selected restore point already contains the unwanted change