Between 4:41 PM PST to 5:10 PM PST on 7/02, DoorDash experienced an incident that affected one of the critical databases. The incident was caused by a bad database query that couldn’t be run efficiently and led to other queries on this database slowing down and eventually timing out.
The issue affected all of DoorDash including Marketplace and Storefront orders and Drive deliveries.
The incident was mitigated after the bad query was reverted and existing query executions terminated.
Majority of the action items following this incident are focused on limiting the potential impact of bad queries on the overall health of our platform.