Mint Uptime Exchange
Flags machines that are likely to fail and gets the repair approved and booked before production stops.
The backend works out what a breakdown would cost and gets sign-off from ops and finance. Then it reserves the parts and assigns a technician. Every repair runs through a 14-step workflow saved in the database, and Server-Sent Events push updates to the frontend.
- Data
- Everything lives in PostgreSQL, from the machines and alerts to the work orders and the audit trail.
- Control
- Logins use JWTs. Each role's permissions come from the database and get checked on every request.
- TypeScript
- Express
- Prisma
- PostgreSQL
- SSE
- Vitest
- machine signalan alert comes in
- risk APIworks out the cost
- approvalops + finance
- inventoryreserved
- work orderscheduled