Inngest Outage on Oct 20 Worsened When Restarts Required the Downed Database

Inngest suffered two disruption windows on October 20, 2026, one from 13:00 to 13:15 UTC and another from 15:20 to 15:35 UTC, affecting event ingestion, scheduling, and execution. The company's report traces both to a connectivity problem between its application infrastructure and database, which stalled database-dependent operations. Event ingestion services ran out of memory and restarted automatically, but their startup required a database connection that was still impaired, so they could not come back online. Inngest's root-cause writeup identifies this as two compounding resilience gaps: memory exhaustion under load and a recovery process dependent on the same resource that had failed. The incident illustrates how an automatic restart can become a second outage when its startup depends on the failing resource.