From silent failures to connections you can watch
Representative scenario — illustrativeA 1,800-student school stitches its systems together by hand: fees settle through a payment gateway, SMS rides a reseller, staff punch a biometric device at the gate, and parents log in with yet another password.
When a parent’s online fee payment “vanished” — debited but never reflected — nobody found out until the reminder went out and the parent arrived furious with a bank statement. The biometric punches were exported to a pen drive and imported by hand, so attendance from the device was always a day behind and sometimes skipped entirely. No one screen ever showed whether a connection was actually working.
Their previous ERP claimed “integrations”, but each one was a fragile one-off with database passwords pasted into a config and no health indicator. When the gateway changed an endpoint the payments silently stopped reconciling, and the school only learned weeks later from angry parents. Six months in, staff had gone back to reconciling gateway settlements against a bank statement by hand — the exact chore the software was bought to remove.
ez.school treats every external service as a managed, monitored connection: payments, SMS/WhatsApp, SSO and biometric devices show last activity and failure state on one screen, so a broken gateway is visible the same day, not a month later. Message sends log per-recipient status, device punches import into live attendance, and anything you build yourselves uses scoped API credentials and webhooks — never a shared database password.
| By hand | Ordinary ERP | ez.school | |
|---|---|---|---|
| A failed gateway | Found via angry parent | Silent, weeks later | Health screen flags it same day |
| Biometric punches | Pen-drive, day behind | Brittle one-off import | Into live attendance directly |
| Credentials | Passwords in a config | Shared DB password | Scoped keys, safely rotated |
| Building your own | Not possible | Not offered | Documented API + webhooks |
- One screenEvery connection’s health
- Same dayFailures surface
- ScopedAPI credentials
Two peaks hit 96% while the library sits at 68% — shift a session to smooth load and defer buying new capacity.
Connection health at a glance — last activity and failure state per service, so a broken gateway is caught before the reminders go out.