Safeguard publishes per-tier system status with live availability, incident history, and subscribable notifications.
Updated continuously · last refresh moments ago
Subscribe with your email to get incident notifications and weekly status digests.
SubscribeNative integrations for enterprise tiers. Route incidents to the channel that owns the on-call rota.
EnterpriseLive availability metrics broken out by deployment tier — Shared, Dedicated, VPC, Sovereign, and Marketplace.
Full incident history for the last 90 days is public. The complete archive is available to customers on request.
Every material incident receives a public RCA within five business days of resolution — timeline, root cause, and remediation actions.
Maintenance windows are published at least 7 days in advance with affected tiers, start time, and expected duration.
Operational status for material sub-processors that affect customer-facing features. Cross-linked from the Trust Center.
Sub-processor status is cross-linked from the Trust Center.
Inference latency in the eu-west region exceeded baseline by 3.2x due to a noisy-neighbour issue on a shared GPU pool. Traffic was failed over to a sibling region; full recovery confirmed at 14:42 UTC.
Full RCAA delay in the upstream vulnerability feed caused new CVE entries to appear with a multi-hour lag. The feed catch-up completed without data loss; affected enrichment jobs were re-run on resolution.
Full RCAPlanned control-plane rolling upgrade. Inference and read-only endpoints remained available throughout; mutation endpoints had a brief queue during the leader transition.
Full RCADuring an active P1, updates land every 30 minutes until resolution.
Initial notification within 15 minutes of detection. Updates every 30 minutes for the duration of an active P1.
Initial notification within one hour of detection. Updates at meaningful state changes through to resolution.
Initial notification within four hours of detection. Single update on resolution plus a written summary.
Acknowledged on the next business day. Bundled into the weekly status digest with the resolution note.
Pick a channel, point it at the right inbox or webhook, and forget about it until something happens — which is exactly when you want it to work.