How Crestfield Digital retired 40 backup scripts
A media company replaced hand-written backup and restore scripts with managed Postgres and infrastructure as code.
Results
Less glue code, more product
- Scripts retired
- Backups, restores, and certificate renewals.
- Recovery point
- min
- Point-in-time recovery for every database.
- Saved each week
- h
- Engineering time back on the roadmap.
The challenge
Crestfield Digital publishes news and podcasts for a dozen regional titles. Its small team maintained 40 shell scripts for database backups, restores, and certificate renewals, each written by someone who had since moved on.
A failed restore test showed that two of the backups had been silently empty for months.
The solution
The team moved its databases to managed Postgres with point-in-time recovery and defined every service and database in a config file checked into the repository. Restores became a single CLI command that anyone on the team can run.
$ cloud db restore newsroom-db --to "2026-03-30 08:15"The results
All 40 scripts are gone. Every database can be restored to any point in the last week, and the team gets six hours a week back for the features their editors had been asking for.
“Managed Postgres with point-in-time recovery means we stopped writing backup scripts and started writing features again.”
