// overall status
Degraded performance
At least one component is running with elevated errors.
Last measured 2026-10-18 13:35 UTC
8 of 9 components operational
// system status
Degraded performance.
Measured from our own job runs every five minutes, not typed in by hand. When something breaks, this page says so on its own.
No active incidents Last measured 2026-10-18 13:35 UTC
// components
Every component, watched.
Each state is derived from the failure rate of that component's own jobs, over a 15-minute and a 60-minute window, and must hold for two readings before it shows here.
Store & catalogue
Reading your products, variants, orders and markets, and writing changes back.
OperationalStorefront & theme
Theme Studio, the GitHub connection behind it, storefront edits and the audits that check them.
OperationalProduct feed
Rendering and pushing your catalogue to Merchant Center, and reading its verdicts back.
OperationalAdvertising
Campaign reporting and the changes the platform makes to your ad accounts.
OperationalTracking & attribution
Collecting click ids and site events, and uploading conversions to the ad platforms.
OperationalCustomer service
Mailbox sync, the agent that drafts and sends replies, and the ticket pages your customers open.
OperationalMarketing Brain
The decision cycle that proposes and dispatches changes across your channels.
Degraded performanceAPI, MCP & webhooks
The published operations, the agent surface and the events we deliver to your systems.
OperationalBackground automation
Everything else that runs on a schedule — billing, reputation, maintenance, recovery.
Operational
// incident history
Public incident history.
Measured incidents from the last 90 days, opened and closed by the same probe as the table above, plus written postmortems for the ones that warranted one.
No incidents in the last 90 days.
This page reflects our public incident history. When there is an incident to report, its postmortem will appear here — root cause, impact, and what changed.
// subscribe
Get notified when status changes.
One email when an incident opens, one when it resolves. No marketing on this list.
By email
One email when an incident opens, one when it resolves. No marketing in this list — ever.
Sent only when status changes
// system status
Honest by default.
This page cannot show green by default: when our measurement stops, every component reads “Not measured” instead of “Operational”.
Public incident history · Postmortems when incidents happen