Vercel outage history

Users attempting to log in to or access pages that use Deployment Protection experienced authentication failures, disrupting access to protected deployments. The issue was traced to the Routing Middleware and began at 4:16 PM UTC. The outage was brief, impacting protected pages for roughly a dozen minutes before a fix was deployed.

Vercel reported elevated error rates for Sandbox in Paris (cdg1) starting 9:16 AM UTC. The issues appeared localized to the Paris Sandbox environment, impacting developers attempting to run tests or deployments there. Other regions were not indicated as affected in the status updates, and users in that area may have experienced intermittent failures or slower responses when using Sandbox features.

From around 1:01 PM UTC to 1:11 PM UTC, Vercel reported elevated build failure rates and deployments returning 500 errors for customers deploying during that window. Some builds progressed but resulted in failed deployments, and users invoking pages or APIs backed by the Edge runtime could also see 500 responses. The impact touched the Build & Deploy pipeline and edge functions, disrupting automated deployments and live edge executions for affected projects.

According to Vercel's status updates, the issue expanded from builds to include edge runtime functions, with initial notices moving from a degraded state to a broader warning before gradually resolving. By 2:16 PM UTC, the status returned to normal as builds and deployments recovered, and guidance to re-deploy Edge runtime functions helped accelerate remediation.

Developers attempting to deploy on Vercel faced delays as builds became stuck in the initializing state, interrupting the deployment workflow. The issue affected the Build & Deploy / Builds component and hindered new deployments for a period around 9:41 PM UTC. Some user reports also indicated issues earlier, around 8:45 PM UTC, signaling impact before official acknowledgment.

Users deploying on Vercel were affected as the Build & Deploy / Builds component experienced an outage driven by elevated errors that triggered deployments unexpectedly. The incident began at Sep 18, 2026 8:36 PM UTC and deployments faced failures or unusual behavior for about fifty minutes, with reports from multiple regions indicating broad impact. Vercel's status page initially showed ongoing investigation as the cause and scope were assessed.

Users relying on Vercel Connect and Passport began experiencing elevated errors starting at 5:46 PM UTC. The issue affected authentication and related workflows for a subset of customers, creating a partial outage in these components. The disruption lasted roughly 1 hour and 35 minutes, with services returning to normal by 7:21 PM UTC.

Status updates showed the issue was identified and a fix was being implemented by 6:16 PM UTC, followed by restoration of Connect and Passport functionality by 7:21 PM UTC.

A latency spike affected multiple Vercel APIs, with the Sandbox API experiencing the most noticeable impact and users seeing slower responses. Over the next hour, latency was observed across the VCR and Blob components as well, reducing the responsiveness of API calls and data retrieval for applications relying on Vercel services. The issue began at 7:16 PM UTC and gradually evolved across the platform.

Developers attempting to deploy on Vercel were affected by failures when deploying to the IAD1 Function regions or using Routing Middleware. The disruption began at 8:01 PM UTC and caused a partial outage for deployments, with the Build & Deploy / Builds component entering a degraded state and deployments failing for affected regions. The issue was regional to IAD1, which likely impacted users attempting builds in that location.

The outage affected users attempting to sign in to Vercel via the CLI, with authentication errors surfacing around Aug 26, 2026 3:11 AM UTC as Vercel began investigating. The Dashboard component also moved into a degraded state and then down by Aug 26, 2026 3:16 AM UTC, signaling a partial outage affecting login flows.

By Aug 26, 2026 4:31 AM UTC the Dashboard had recovered and the incident was considered resolved. Vercel's status updates described ongoing investigation and remediation to restore login functionality.

Developers using Vercel's Build & Deploy / Builds experienced elevated initialization times beginning 7:31 PM UTC. The issue prompted ongoing updates as engineers implemented a fix, with 8:01 PM UTC noting that a fix had been deployed and monitoring was underway. The service returned to normal by 8:06 PM UTC, after roughly 35 minutes of degraded startup times for builds.