Briefing
The host that runs the storefront also runs a historical order backfill, and for the last ten minutes it has been pinned at full CPU.
Storefront p95 latency has doubled on the dashboards, although requests still succeed.
Operations has approved pausing the backfill until its input data is repaired, and nothing customer-facing is to be touched.
The first guess in the channel was a traffic spike on the storefront.
The fix is yours to find: hints, the recovery checks and the debrief unlock inside the incident.