From 20:04 to 21:12 CEST a subset of merchant-portal page loads and collection-data requests were extremely slow (30 seconds to several minutes in the worst window 20:38–20:48 CEST); requests completed without errors and no data was lost. Root cause: a throughput ceiling in the network path between our edge platform and database, reached during an unusually large ingestion traffic surge. We deployed a transport-protocol fix at 21:09 CEST and latency returned to normal within three minutes. Follow-up capacity and alerting improvements are in progress.