> ## Documentation Index
> Fetch the complete documentation index at: https://grandcentral.backbase.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Data protection and backup

> How the Backbase Agent Platform protects PostgreSQL data with automated backups and point-in-time restore

The Backbase Agent Platform stores agent and knowledge data in Azure Database for PostgreSQL Flexible Server. The platform manages backup and retention. You don't configure backup policies.

## At a glance

The following table summarizes platform backup standards.

| Topic            | Platform standard                                                                                                                                                  |
| ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Backup mechanism | Azure-managed automated backups with point-in-time restore (PITR)                                                                                                  |
| Retention        | Development: 7 days; test and staging: 14 days; production: 35 days                                                                                                |
| Your role        | Provision databases through self-service; backup is platform-managed                                                                                               |
| Data recovery    | Submit a [Jira request to the platform team](https://backbase.atlassian.net/jira/software/c/projects/ECOPFM/form/3448?from=directory); restore is not self-service |

## How backups work

Azure automatically protects each PostgreSQL Flexible Server with:

* Daily snapshot backups (full initial snapshot, then differential snapshots when data changes)
* Continuous transaction log (WAL) archiving for point-in-time restore within the retention window

Backups cover all databases on a runtime's PostgreSQL server. The platform can restore to a specific point in time within the retention period for that environment.

<Info>
  Point-in-time restore creates a separate restored instance for recovery workflows. It does not overwrite your live database in place.
</Info>

## Recovery objectives

The following table lists recovery objectives. These are planning targets, not customer-specific contractual SLAs unless agreed separately with your organization.

| Metric                         | Planning target               | Notes                                                                              |
| ------------------------------ | ----------------------------- | ---------------------------------------------------------------------------------- |
| Recovery Point Objective (RPO) | Up to 15 minutes              | Depends on transaction volume; Azure archives WAL on a workload-dependent schedule |
| Recovery Time Objective (RTO)  | Best effort, minutes to hours | Depends on database size and how far back you need to restore                      |

## Backup vs disaster recovery

The following table compares backup protection with disaster recovery and archival. Standard backups stay in the same Azure region unless geo-redundant backup is agreed at server creation.

| Scenario                               | Protection                                                                                                        |
| -------------------------------------- | ----------------------------------------------------------------------------------------------------------------- |
| Accidental data change or deletion     | Automated PITR within the retention window                                                                        |
| Azure region outage                    | Disaster recovery procedures that are separate from backup. The platform team manages regional disaster recovery. |
| Long-term archival (more than 35 days) | The standard platform backup window does not cover long-term archival                                             |

## What you need to know

* You don't configure backup retention. The platform sets retention per environment class.
* You provision databases through your `applications-live` repository and the `infra-resources` self-service flow.
* If you need a data restore, submit a [Jira request to the platform team](https://backbase.atlassian.net/jira/software/c/projects/ECOPFM/form/3448?from=directory). Include the runtime, database name, and the point in time you need to recover to. The platform team performs operational restore steps.

## Related documentation

For more information, see:

* [Backbase Agent Platform overview](/agentic-ai/backbase-agent-platform)
* [Technology stack](/agentic-ai/architecture/technology-stack)
* [Self service](/agentic-ai/get-started/self-service)