ISSUE SUMMARY

On Sunday 18 January 2018, Google Compute Engine networking experienced a network programming failure. The two impacts of this incident included the autoscaler not scaling instance groups, as well as migrated and newly-created VMs not communicating with VMs in other zones for a duration of up to 93 minutes. We apologize for the impact this event had on your applications and projects, and we will carefully investigate the causes and implement measures to prevent recurrences.

DETAILED DESCRIPTION OF IMPACT

On Sunday 18 January 2018, Google Compute Engine network provisioning updates failed in the following zones: europe-west3-a for 34 minutes (09:52 AM to 10:21 AM PT) us-central1-b for 79 minutes (09:57 AM to 11:16 AM PT) asia-northeast1-a for 93 minutes (09:53 AM to 11:26 AM PT)

Propagation of Google Compute Engine networking configuration for newly created and migrated VMs is handled by two components. The first is responsible for providing a complete list of VM’s, networks, firewall rules, and scaling decisions. The second component provides a stream of updates for the components in a specific zone.

During the affected period, the first component failed to return data. VMs in the affected zones were unable to communicate with newly-created or migrated VMs in another zone in the same private GCE network. VMs in the same zone were unaffected because they are updated by the streaming component.

The autoscaler service also relies upon data from the failed first component to scale instance groups; without updates from that component, it could not make scaling decisions for the affected zones.

ROOT CAUSE

A stuck process failed to provide updates to the Compute Engine control plane. Automatic failover was unable to force-stop the process, and required manual failover to restore normal operation.

REMEDIATION AND PREVENTION

The engineering team was alerted when the propagation of network configuration information stalled. They manually failed over to the replacement task to restore normal operation of the data persistence layer.

To prevent another occurrence of this incident, we are taking the following actions: We still stop VM migrations if the configuration data is stale. Modify the data persistence layer to re-resolve their peers during long-running processes, to allow failover to replacement tasks.

The issue with Google Cloud Networking connectivity has been resolved for all affected zones in europe-west3, us-central1, and asia-northeast1 as of 11:26am US/Pacific. We will conduct an internal investigation of this issue and make appropriate improvements to our systems to help prevent or minimize future recurrence. We will provide a more detailed analysis of this incident once we have completed our internal investigation.

We are investigating an issue with Google Cloud Networking affecting connectivity in us-central1 and europe-west3. We will provide more information by 12:15pm US/Pacific.