Resilience & recovery

Targets we drill against, not slogans

A platform that touches money has to fail well. We publish recovery targets per surface tier, the cadence we drill them at, and the outcome of the most recent drill. Anything missed lands on the trust ledger the same day.

Ledger RTO

≤ 15 min

RPO 0 · synchronous

Payouts RPO

≤ 60 s

HSM + dual bank rails

Drill cadence

Monthly

Per tier, rotating surfaces

Last full game day

2026-06-08

Findings on /tabletops

Recovery targets by tier

Tier	Surfaces	RTO	RPO	Last drill	Outcome
Tier 0 · Ledger	Journal, Merkle log, attestations	≤ 15 min	0 (synchronous replication)	2026-06-08	Recovered in 9m12s; root recomputed and matched.
Tier 1 · Payouts	Treasury queue, HSM, bank rails	≤ 30 min	≤ 60 s	2026-05-25	Failed over to secondary EU corridor; payouts resumed in 22m.
Tier 1 · Platform APIs	Workspace API, dashboards	≤ 30 min	≤ 60 s	2026-05-11	Read traffic shifted in 4m; full write parity in 18m.
Tier 2 · Recon & pacing	Recon engine, Pacer	≤ 2 h	≤ 5 min	2026-04-20	Recovered in 51m; backfill caught up within the hour.
Tier 3 · Reporting	Board pack, exports, analytics	≤ 8 h	≤ 1 h	2026-04-06	Recovered in 1h44m; warehouse rebuild from snapshot.

Regional posture

Primary region EU-West with hot standby in EU-North; failover is automated and tested monthly.
HK region is hot-active for APAC payouts with paired EU read replica for ledger continuity.
Cross-region traffic shifting via DNS + service mesh; tested under load on every game day.

If we miss a target

Customer notification within 1 hour of detection; incident report within 5 business days.
Automatic SLA credits where the breach maps to /service-levels; visible in /service-credits.
Root cause and remediation owner posted to /trust-ledger and /transparency.