GCP services in the us-central1 region experienced intermittent network packet loss for traffic entering Google's network in the Chicago region, from 13:10 to 14:12. Traffic entering Google's network in other regions was not affected. Our engineers are busy collecting data, but their preliminary assessment is that approximately 5% of total traffic to the us-central1 region was affected during the event, which was caused by two separate network router failures. Other GCP regions were not materially affected by the event. We expect to continue to investigate both the cause and the impact of the event, and will update affected customers with more information as well as a corrective plan within 3 business days.

We are experiencing a networking issue with Google Compute Engine instances in us-central zones beginning at Wednesday, 2017-03-15 01:00 PM US/Pacific.