Understanding retention in Endpoint Backup

Overview

Endpoint Backup uses a Grandfather-Father-Son (GFS) retention model to manage your recovery points efficiently. Retention runs automatically each night, pruning older backups while maintaining a healthy chain for restore integrity. This document explains how retention works, how to interpret calendar-based rollups, and what to expect from each retention tier.

How the Retention Process Works

The retention job runs nightly as part of Datto's scheduled operations. During each run:

  • The data in older recovery points are deleted based on the retention policy.
  • This process is similar to data "pruning" because recent backups are kept in greater detail, while older backups are summarized.
  • Backup integrity is preserved. No recovery points are removed if they are required to maintain a valid chain.

Retention is policy-driven and can be configured or adjusted at any time with the Policies section in UniView.

NOTE  Retention still runs even for paused agents. Custom retention settings should be reviewed if agents are no longer actively backing up.

Understanding the grandfather-father-son (GFS) strategy

GFS is a tiered backup rotation strategy:

  • Son – Daily backups (short-term recovery points)
  • Father – Weekly backups (mid-term recovery points)
  • Grandfather – Monthly backups (long-term archival points)

Backups roll up over time as follows:

  • Intra-Dailies → Dailies
  • Dailies → Weeklies
  • Weeklies → Monthlies

This structured pruning ensures recent restore points are available in high resolution, while older backups are retained at broader intervals to save storage.

Calendar-based retention vs. exact date retention

Endpoint Backup uses calendar-based retention, which is different from rolling, “X-days ago” logic.

For example, a 3-year retention policy means:

  • You won’t have a recovery point from exactly 36 months ago.
  • Instead, you’ll have structured restore points per policy: daily, weekly, and monthly.

Example

  • Today: August 2025
  • Retention: 3 Years
  • Oldest backup available: December 2022

This approach helps with predictable long-term archival and version control while reducing storage overhead.

Retention by tier

Each predefined retention plan includes the following GFS tiers:

Retention Plan Intra-Dailies Dailies Weeklies Monthlies
90 Days 14 Days 7 Days 1 Week 2 Months

1 Year

14 Days 7 Days 1 Week 11 Months
3 Years 14 Days 7 Days 1 Week 11 Months
7 Years 14 Days 7 Days 1 Week 11 Months

Each tier defines how far back a particular granularity of backups is retained.

NOTE  All backups are stored incrementally. Each point contains only the changes since the previous one, which affects how much space can be freed during pruning.

Tier breakdown

Intra-Dailies → Dailies

  • Multiple backups per day for N days (e.g., 14).
  • Pruned to the last backup of the day, depending on retention time.

Dailies → Weeklies

  • After ~7 days, daily backups roll up into one weekly point (the last backup of the week).

Weeklies → Monthlies

  • After 4 weeks, weekly points are reduced to one monthly point (the end of the month).

NOTE  Retention never removes data required to preserve the chain’s integrity.

Data availability and recovery impact

When retention removes older backups, it only keeps the data available at the time that backup was taken. Files deleted or changed on the source system before a retained snapshot was captured will not be available for recovery.

For example:

  • If a file was deleted on the endpoint in June, and your only retained backup is from December, that file will not be in the December backup.

Storage optimization & cloud retention

  • The cloud retention process can delete multiple points in a single run.
  • Large datasets may require additional time for pruning and consolidation.
  • Backup chains are optimized to avoid duplication across restore points.

Summary: Why we use calendar-based GFS retention

  • Maintains frequent recovery points for recent activity.
  • Reduces granularity over time to balance storage and restore options.
  • Ensures predictable restore history that meets compliance and storage efficiency goals.