INTERNAL DESK MEMO // CONFIDENTIAL

EBA Clearing RT1 Latency & Peak Hour Queue Analytics

Date: May 2026 • Desk: Liquidity Operations & Settlement Telemetry • Ref: LN-RT1-042

Desk Observation: EBA Clearing RT1 settlement latency metrics exhibit significant performance degradation during peak liquidity sweeps. This brief presents raw telemetry capturing empirical 99.9th percentile latency outliers and queuing limits, bypassing the institution's generalized public averages.

RT1 99.9th Percentile Latency (Peak Retail Sweep Period) 1,842 ms Threshold SLA target: < 500 ms

Empirical Latency telemetry

During the 16:30-17:00 CET daily retail sweep window, message processing buffers at direct participants encounter memory congestion. telemetries indicate a clear divergence between retail transfer finality and atomic on-chain settlement rails. While a 1.8-second delay passes unnoticed in general retail, it introduces an asymmetric settlement gap for algorithmic market makers bridging DLT cash legs to legacy API bridges.

WINDOW (CET) AVG LATENCY 99.9th PERCENTILE QUEUE COUNT
09:00 - 10:00 124 ms 340 ms Low
13:00 - 14:00 245 ms 980 ms Medium
16:30 - 17:00 612 ms 1,842 ms High (Congested)

Relying on legacy message queues as cash rails for programmatic trading exposes desks to toxic arbitrage. When compared with TARGET2 weekend liquidity bottlenecks, the data demonstrates that asynchronous API bridging leads to settlement imbalances.

[UNRECONCILED RAW FEED MATCH - FEED_RT1_DEV_B]
03:14:22.102 | SEQ_ID: 981123 | EBA_REF: RT1-SCT-012 | STATUS: PENDING_QUEUE
03:15:02.404 | SEQ_ID: 981124 | EBA_REF: RT1-SCT-013 | STATUS: TIMEOUT (RELOADED)
03:15:45.912 | SEQ_ID: 981125 | EBA_REF: RT1-SCT-014 | STATUS: CLEAR_SUCCESS

* TELEMETRY NOTICE: Conflicting interpretations across ECB primary reports vs private RT1 endpoint logs prevent stable convergence mapping. Telemetry above is preliminary and subject to validation under high stress cycles.
DESK DISCLAIMER & LIMITS: Data is compiled via passive port-sniffing at selected target participant endpoints. Performance varies by participant gateway processing speeds, local middleware configurations, and hardware acceleration constraints. EBA Clearing does not provide direct millisecond-level telemetry to non-direct participants.