Incident response
How we handle security incidents.
Last updated 09 May 2026
Our incident response policy:
Detection
· WATCHDOG CRON (15-min cadence): infrastructure + application health
· GUARDIAN CRON (daily): PDPA + cross-tenant isolation
· SENTINEL CRON (hourly): output quality + safety
· Customer reports: security@onset.my, 24-hour acknowledgement target
Classification (severity)
· SEV1 — major outage, data exposure, customer-impacting bug spreading
· SEV2 — partial outage, degraded performance
· SEV3 — informational degradation, no customer impact
· SEV4 — internal observation
Response
· On detection, an incident is logged at /admin/incidents.
· A war-room is opened in Telegram with the on-call.
· Public-facing updates land at /status within 15 minutes for SEV1/SEV2.
· Affected customers are emailed within 24 hours for any incident that touched their data.
Post-incident
· A blameless post-mortem is published internally within 5 working days.
· Customer-visible incidents have a public post-mortem within 14 days.
· SLA credits are applied automatically per /sla.
PDPA security breach: notified to NDPC within 72 hours of confirmation, per the PDPA Amendment 2024.
24/7 contact: security@onset.my (PGP key on request).
