Set up scheduled, off-site, restore-tested Postgres backups with one script and a cron job.
Cloud & DevOpsPDF · 7 pages· v1.0
4.7A complete system for backing up a PostgreSQL database off-server to S3-compatible object storage (AWS S3 or DigitalOcean Spaces) automatically, with retention, encryption-at-rest by the provider, and a tested restore procedure. Many teams take backups but never verify them. This guide fixes both halves: it gives you a robust pg_dump backup script that uploads to object storage, a cron schedule, sensible retention/cleanup using lifecycle rules, and — crucially — a step-by-step restore drill so you know your backups actually work before you need them. You'll configure the AWS CLI (which works with both S3 and Spaces via a custom endpoint), create a least-privilege bucket and access key, write the backup script with safe defaults and logging, and schedule it. The guide explains the difference between logical (pg_dump) and physical backups, when each matters, and how to handle large databases. After this guide you'll have hands-off nightly backups landing in a bucket, automatic pruning of old ones, alerting on failure, and a restore runbook you've actually executed once. Works for self-hosted Postgres on a VPS or in Docker.
Both. Spaces is S3-compatible; the guide shows the single --endpoint-url change that makes the AWS CLI talk to Spaces.
This guide uses logical backups (pg_dump), which are portable and simple. It explains when you'd reach for physical/WAL-based backups instead.
Via the bucket's lifecycle/expiration rules plus an optional script-side prune. The guide configures retention so storage doesn't grow forever.
You run the included restore drill. The guide walks you through restoring into a scratch database and verifying row counts.
Read the full refund policy and trust & safety terms.