App rescue
Production rescue
Two live apps for one client, taken over and made stable, one right after its production database was wiped.
Overview
One client, two existing apps built by the founder, March to September 2026. We took both over and made them stable: an appraisal pipeline built on FastAPI and React, and a Next.js app we took over right after its production database was accidentally wiped.
The appraisal pipeline
A FastAPI and React app that ran a pipeline in phases. Startup was not reliable. After a server restart it created a blank session instead of loading the saved state. Phases that had already finished ran again instead of loading their cached results. Errors failed silently, and there were no error boundaries to catch them.
The Next.js app
A Next.js and PostgreSQL app, taken over right after its production database was accidentally wiped. Once the data was recovered, our work started: protecting the database, cleaning up what had been recovered, and fixing its reliability bugs.
What changed
- Startup fixed, so the pipeline app comes up reliably
- Restarts load the saved state instead of opening a blank session
- Finished phases load cached results instead of running again
- Silent errors fixed and error boundaries added
- A guard now protects the production database of the Next.js app
- Recovered data cleaned up and reliability bugs fixed
Want the engineering detail, the problem, the build, and the hard production parts? Read the full case study