**Summary** Some customers may experience slowness while using Google Drive. **Description** We are experiencing an issue with Google Drive beginning Tuesday, 2...

What caused it: Google Drive's root cause analysis

# Incident Report Summary On Tuesday, 18 August 2026 at 16:58 US/Pacific, an incident affecting Google Drive occurred for a duration of 12 hours and 42 minutes. Customers located in Australia, Philippines and New Zealand experienced severe latency, sporadic loading failures, and errors when accessing Google Drive.  To our valued Google Workspace customers, we are deeply sorry for the service interruption you experienced. We are taking immediate and decisive action to address the issue and strengthen our systems against future incidents. Other Google Workspace services were not affected. Root Cause Google utilizes a multi-tiered global load balancing and network routing architecture to serve traffic efficiently. To ensure fair use and prevent congestion, bandwidth allocation is managed by attributing traffic to specific downstream services. A misconfiguration in the attribution of traffic caused the bandwidth usage to be incorrectly attributed to the frontend network load balancing tier itself, rather than to the Drive service. The sequence of events leading to customer impact was as follows: * Traffic from the regional load balancing tier was incorrectly attributed to the load balancing infrastructure rather than the Drive service. * Unrelated traffic caused network bandwidth usage attributed to the load balancing infrastructure in this region to exceed its quota. * Network bandwidth management systems applied throttling to this traffic to enforce quotas, causing severe network congestion affecting Drive traffic in this region. * This congestion led to increased latency, slowness, and error rates for users accessing Drive in the affected regions. Remediation and Prevention Google engineers were alerted to the issue via automated monitoring and an influx of customer support cases, with initial incident response teams engaged by 17:21 US/Pacific on 18 August , 2026\. Engineers initially attempted to alleviate the congestion by draining regional frontend traffic to an alternate geographic location, but this proved ineffective and was reverted. The team then mitigated the issue by migrating regional traffic to an upgraded internal network routing architecture, which correctly set the traffic attribution for Drive. Normal operations were fully restored at 05:40 once 100% of the regional traffic was migrated to the upgraded network architecture. To prevent a recurrence of this issue, our engineering teams are undertaking the following structural preventative act...

Read the full postmortem from Google Drive