Check integration and data-job health

Control principle

A connection is not reliable until failure is visible, owned, and recoverable.

Automations, imports, exports, syncs, or reports can fail quietly until staff or constituents notice missing information.

Use it in the active service

Start with these controls

1. List each data movement, schedule, source, destination, business purpose, owner, and expected completion signal.
2. Define freshness, completeness, exception, retry, reconciliation, and alert thresholds that reflect real service impact.
3. Test an approved failure and recovery path without exposing production secrets or protected data.
4. Record recurring defects, manual workarounds, vendor dependencies, and the next reliability decision.

Pause point

Do not let operating pressure erase nonprofit ownership

A green technical status may not prove correct records or outcomes. Reconcile important work with accountable business owners.