Alert Configuration

A QC Alert Configuration turns data into action. An inspection records that a defect rate crossed 8%; without alerts, that fact sits in the dashboard until someone happens to look. With an alert wired up, the production manager gets a notification within seconds and the line can pause before another 50 garments come off with the same flaw.

If parameters, IQ templates, and defect categories decide what data to collect, alerts decide what to do with it. The two halves work together — captured data without alerts is filing; alerts without captured data fire on nothing.

Getting there

Awesome bar: Cmd/Ctrl + K → type qc alert configuration → pick QC Alert Configuration List

Click path: Desk → GarmentFlow → Quality Control → Setup card → QC Alert Configuration

Each alert is its own record. Most factories end up with 10–20 of them, each tuned for a specific kind of signal.

What an alert is built from

The fields:

• Alert Name — descriptive. "Critical defect found on Final QC." "Three consecutive in-line fails on stitching."
• Is Active — disable without deleting.
• Trigger Event — one of five:
• Defect Rate Threshold Exceeded — defect rate on an inspection crosses your threshold.
• Critical Defect Found — any defect with severity Critical is logged.
• Repeated Defect Pattern — same defect category found N times in a configured time period.
• NCR Raised — a Non-Conformance Report is created.
• Inspection Failed — any inspection lands on Rejected or Hold for Review.
• Item Group — narrow the alert to a class of items. Optional.
• Defect Category — narrow to a specific category. Optional.
• Threshold Type and Value — Defect Rate %, Count per Batch, or Consecutive Failures, plus the numeric threshold.
• Time Period — Per Batch, Daily, Weekly, Monthly. Defines the window for counting.
• Alert Method — Email, System Notification, SMS, or any combination.
• Severity Level — Info, Warning, Critical. Drives notification urgency and dashboard prominence.
• Recipients — child table linking users or roles to be notified.
• Message Template — the subject and body of the notification, with placeholders for batch ID, defect category, value, etc.

When alerts work, when they don't

The trap with alerts is wiring too many at the wrong thresholds. The signal-to-noise ratio matters more than coverage. Three principles:

1. Tune for action, not awareness. An alert that fires "for information" gets ignored within a week. Every alert should trigger a specific action by a specific person.
2. Use Severity deliberately. Reserve Critical for things that genuinely need someone to stop what they're doing. Warning for "act soon." Info for the rare case where a logged record matters without urgency.
3. Watch the noise. If an alert fires more than a few times a week without action, the threshold is too low or the audience is wrong. Re-tune; don't tolerate.

A working example

A starter set of alerts for a mid-size factory:

• Critical defect on Final QC → Trigger: Critical Defect Found. Item Group: Tops. Severity: Critical. Recipients: Quality Manager, Production Manager, Plant Head. Method: System Notification + Email. Acts like a stop-the-line signal.
• Final QC defect rate over 5% → Trigger: Defect Rate Threshold. Threshold: 5%. Period: Per Batch. Severity: Warning. Recipients: Quality Manager. Method: System Notification. Flags batches that need a closer look before shipping.
• Same stitching defect three times in a day → Trigger: Repeated Defect Pattern. Defect Category: Stitching → Skipped stitch. Threshold: 3 consecutive failures. Period: Daily. Severity: Warning. Recipients: Production Manager + Line Supervisor for the affected unit. Catches a drifting machine.
• In-line inspection failed → Trigger: Inspection Failed. Severity: Warning. Recipients: Line Supervisor + Quality Manager. Method: System Notification. The pause-the-operation signal.
• NCR raised → Trigger: NCR Raised. Severity: Critical. Recipients: Quality Manager + Plant Head. Method: Email. The formal escalation path.

Five alerts. Each one drives a specific action by a specific person. Each one fires rarely enough that recipients actually pay attention.

Image: A QC Alert Configuration record for "Same stitching defect three times in a day," with trigger conditions, recipients, and message template visible.

What to do next

That closes setup. The data is being captured, the alerts are wired. The next step is reading what comes out — head to Dashboards.