On September 22, 2025, beginning at 17:30 UTC, we experienced an issue that led to performance degradation for a subset of customers using our NodeBalancer product.

The root cause was sustained high CPU utilization on the shared compute infrastructure that underpins the NodeBalancer service. Because this infrastructure is shared among customers, unusually high demand from a few customers pushed CPU usage close to 100%, resulting in degraded performance for others on the same hardware.

This issue impacted multiple regions, including LAX3, LON4, ORD2, PAR3, and OSA1. Our engineering teams quickly traced the elevated CPU usage to overloaded NodeBalancers and worked to stabilize the environment. Mitigation efforts were prioritized to restore performance, with full resolution completed by 15:49 UTC on October 6, 2025.

We are taking proactive steps to reduce the likelihood of similar incidents in the future. This includes a range of improvements aimed at enhancing overall system efficiency and stability. We remain committed to transparency and reliability, and we appreciate your patience as we continue to strengthen our platform and deliver consistent service.

This summary provides an overview of our current understanding of the incident given the information available. Our investigation is ongoing and any information herein is subject to change.