Tweet by grinich

July 17, 2026

Yesterday, WorkOS experienced a service disruption that caused elevated latency, timeouts, and authentication failures for many customers. The incident was triggered by an unusual increase in session-creation traffic, but the root cause was ours: a latent transaction bug and insufficient workload isolation allowed non-critical analytics work to exhaust a shared database connection pool. We fixed the defect, restored service, and are moving analytics out of the critical authentication path, adding stronger overload protection, and improving our ability to detect and contain failures faster. Authentication infrastructure has to be dependable. We fell short of that standard and we’re sorry to our customers and their users who were affected. Full postmortem: https://t.co/qqBzezIUdI

Author
grinich
Date
July 17, 2026
Open on X