Wed Sandbox Drift Audit 2026-05-06 — recommend refresh planning, not auto-refresh
Decision
Sandbox drift threshold breached (13 actionable components > 10) but recommend NEW planning task before any refresh — auto-refresh would wipe 7 in-flight discount/quote-engine customs (Discount_Details__c, Offer_Master__c, Offer_Master_Data__c, Quote_Applied_Discounts__c, Voucher_Details__c, Project_Development__c, CP_BillingChekclist_records__c) plus GST_Conversion__mdt. ARJ-48 closes with finding posted; follow-up task created for refresh planning (NOT execution).
Rationale
Raw drift = 178 (139 prod-only + 39 sandbox-only) but most are platform-companion artifacts (__History/__Share/__Feed/__ChangeEvent following root objects) or Revenue/Product Cloud standard objects not enabled in sandbox edition. Net actionable: 5 prod-only Runwal customs (Booking_Cancel__c, CP_Sourcing_dashboard__c, Document_Checklist__c, Parking__c, Unit_Event_History__c — sandbox is stale) + 7 sandbox-only Runwal in-flight customs + 1 sandbox-only mdt = 13. Refreshing fullcopy without first confirming source-control coverage of the 7 in-flight customs would destroy unsync’d dev work — a Three-Org-Discipline violation worse than the staleness it would fix. Field-level diff sampled only 10 of 2,052 common objects, so a Phase 2 targeted field-diff on top Runwal objects is also needed before refresh sign-off.
Alternatives Rejected
A. Auto-refresh fullcopy now — REJECTED: destroys 7 in-flight customs without source-control verification. B. Ignore the drift, defer to next Wednesday — REJECTED: 5 prod-only customs are real Runwal business objects (Parking, Document_Checklist, Booking_Cancel, Unit_Event_History, CP_Sourcing_dashboard) that should exist in sandbox; staleness compounds. C. Recommend planning task + create follow-up — CHOSEN: surfaces threshold breach to CTO with classified finding, defers destructive action to a planning session that confirms source-control coverage and decides refresh-vs-forward-deploy of in-flight work.
Outcome
Pending