Digital Channel SLA Calculation Discrepancy in EU-West BYOC

Need some troubleshooting help with a variance in Service Level calculations for Digital Messaging within the EU-West BYOC environment.

The Performance Dashboard reports a 92% SLA achievement for the primary support queue, yet the underlying Architect flow logic indicates that 15% of interactions exceeded the 30-second response threshold. The discrepancy appears isolated to chat sessions initiated via the Web Widget SDK v2.4.1.

Is there a known latency offset between the widget event timestamp and the Genesys Cloud ingestion timestamp that would cause this metric drift?