4.4 Escalation SLAs

Hope is not a reliable strategy. When a line stops, the clock directly impacts the factory’s P&L. An “Escalation SLA” (Service Level Agreement) is a programmed set of rules that governs the human response to downtime. It aims to remove emotion and ambiguity: if the line is down, the system should summon help automatically.

The SLA matrix by event type

Distinct workflows must be defined for different failure modes. A specialized machine error typically requires a different responder than a cardboard shortage.

Event Class	Trigger Condition	Primary Responder (Tier 1)	Response SLA (Max Time)
Machine Down	Machine State = Error > 5 Mins OR Operator “Maintenance Call” Button.	Maintenance Technician	10 Minutes
Quality Stop	Consecutive Yield < 90% OR Critical Test Fail (e.g. Hipot).	Process / Quality Engineer	15 Minutes
Material Starved	”Feeder Low” Warning OR Operator “Material Call” Button.	Water Spider / Logistics	5 Minutes
Traceability Gap	System Interlock: “Genealogy Link Missing” or “Profile Mismatch”.	MES Super User / Quality Admin	10 Minutes
IT/Network	Server Ping Fail OR HMI Freeze.	IT Support (L1)	5 Minutes

Logic:

• When an event occurs, the system starts a timer (T=0).
• When a responder scans their badge at the station within the defined SLA, the timer pauses, and the state changes to “In Progress.”
• When the timer exceeds the SLA, the system should trigger an escalation to Tier 2.

The escalation hierarchy (automatic promotion)

The system objectively escalates based on time. If the problem is not solved within the defined SLA, the notification moves up the management chain.

Tier 1: the tactical response (t + 0 min)

• Who: Line Technician, Line Lead, Water Spider.
• Notification: Andon Board (Yellow/Red), Smart Watch/Pager.
• Goal: Quick fix / Reset.

Tier 2: the engineering response (t + 15 min)

• Who: Process Engineer, Maintenance Supervisor, Quality Manager.
• Trigger: Tier 1 failed to resolve the issue (or failed to acknowledge it) within 15 minutes.
• Notification: SMS / Mobile Push Notification.
• Goal: Root cause analysis, advanced troubleshooting.

Tier 3: the executive response (t + 60 min)

• Who: Plant Manager, Director of Operations.
• Trigger: Line down > 1 Hour.
• Notification: Email / Urgent SMS.
• Goal: Resource reallocation, overtime authorization, customer impact assessment.

Traceability gap protocol (special handling)

A Traceability Gap (e.g. “Parent unit passed, but the embedded Child component has no scan record”) is not a typical machine fault; it represents a Compliance Breach.

• Severity: Critical.
• Action: Initiate an Immediate Hard Stop of the affected line section.
• Responder: This must be addressed by a System Admin or Quality Manager. Operators should not have permissions to override genealogy errors.
• Resolution: A manual data patch (if physical proof exists to justify it) or Scrapping the Unit.

Closure rules: the “confession”

Closing a ticket is a critical data entry event. The system should require the responder to categorize the failure before the line allows a restart.

Mandatory fields

• Root Cause Code: A standard tree must be selected from (e.g. M_Motor_Fail, Q_Solder_Bridge). Using an open-ended “Other” category should be avoided.
• Action Taken: A brief text description must be provided (e.g. “Replaced sensor X”).
• Duration: This should be auto-calculated by the system (Time_Closed - Time_Opened).

The “micro-stop” filter

• Scenario: The machine experiences an error, but the operator resets it almost immediately (Duration < 2 minutes).
• Logic: The system should not demand a manual entry for these very brief events. It should auto-log them as a “System_Microstop”.
• Review: When the count of micro-stops exceeds 10 per hour, the system should trigger a Tier 2 Alert to investigate the chronic issue.

Verification scan

• Rule: A Maintenance ticket should remain open until the machine successfully produces 1 Good Unit.
• Logic:
  1. The Technician fixes the machine.
  2. The Technician updates the ticket status to “Verify”.
  3. The Operator runs a unit through the process.
  4. When the result is a “Pass,” the ticket automatically closes.

Final Checkout: Escalation SLAs

Category	Metric / Control	Threshold / Rule
Response	Time-to-Ack	Ensure Tier 1 acknowledges (e.g. Scan Badge) within the defined SLA (5-15 mins).
Escalation	Auto-Trigger	When Time > 15m (unresolved), the system should auto-send an SMS to Tier 2.
Data	Taxonomy	Ticket closure requires a specific Root Cause Code (Avoid “Misc” or “Other”).
Integrity	Verification	Maintenance tickets ideally require 1 successful “Pass” cycle to close.
Compliance	Traceability	Genealogy gaps should require a Manager-level override to clear.
Visibility	Andon	Ensure the Andon Board reflects the current Escalation Level (e.g. Flashing Speed increases).
Analysis	Pareto	Conduct a weekly review of the top offenders based on Sum(Duration) by Reason Code.