Is Vercel down?

Understanding Vercel availability

Vercel is a cloud and developer platform for deploying and serving web projects. A reported incident on Vercel’s official status page is the strongest initial signal of a broader platform problem, but it does not confirm that every project, domain, or deployment is affected.

A Vercel-wide issue becomes more plausible when unrelated projects, deployments, dashboard actions, or other platform functions fail at approximately the same time. If one project or custom domain fails while the status page shows no matching incident, investigate that deployment’s build output, runtime behavior, environment configuration, DNS, certificate, or redirect settings.

The message shown in a browser can also describe a failure on the path between your device and the requested site. Record the exact error, affected hostname, time of failure, whether another Vercel project works, and whether the same request succeeds from another device or network before changing configuration.

Troubleshooting Vercel problems

Check for a platform incident

Open Vercel’s official status page and compare any reported incident with the feature that is failing. If the incident matches, note the affected deployment or project and the time of failure, avoid unrelated configuration changes, and retry when the incident information indicates recovery. If no relevant incident is listed, continue with project, domain, and local checks.

When one project or deployment fails

Inspect the affected deployment’s build and runtime output. Confirm that the expected deployment is connected to the requested hostname and compare it with a known-working deployment when possible. A failure isolated to one project points more strongly to application code, build settings, environment configuration, or domain setup than to a platform-wide outage.

When the browser shows an HTTP error

• HTTP 500: Investigate the server-side application or deployment failure and review recent code and environment changes. See [HTTP 500 Internal Server Error](/guides/http-500-internal-server-error).
• HTTP 502: Check whether an upstream service or deployment process returned an invalid or unavailable response. See [HTTP 502 Bad Gateway](/guides/http-502-bad-gateway).
• HTTP 503: Compare the timing with Vercel’s status page and inspect the affected deployment. See [HTTP 503 Service Unavailable](/guides/http-503-service-unavailable).
• HTTP 504: Look for a request or upstream operation that is taking too long. See [HTTP 504 Gateway Timeout](/guides/http-504-gateway-timeout).

When a custom domain has DNS, SSL, or redirect problems

A custom-domain error does not by itself indicate that Vercel is unavailable. Check recent DNS changes, the records for the requested hostname, and whether the hostname and certificate configuration match. For a redirect loop, test the intended HTTP and HTTPS addresses and review the project’s redirect rules.

Useful related diagnostics include [DNS_PROBE_FINISHED_NXDOMAIN](/guides/dns-probe-finished-nxdomain), [DNS_PROBE_FINISHED_BAD_CONFIG](/guides/dns-probe-finished-bad-config), [ERR_CERT_COMMON_NAME_INVALID](/guides/err-cert-common-name-invalid), and [ERR_TOO_MANY_REDIRECTS](/guides/err-too-many-redirects).

When the issue affects one browser, device, or network

Try the URL in a private window, temporarily disable extensions, and test from another device or network. If it works elsewhere, investigate browser state, local DNS caching, or the original network route before changing the deployment. Connection-refused and connection-timeout messages may originate from the client path or a specific origin request rather than from Vercel as a whole. See [ERR_CONNECTION_REFUSED](/guides/err-connection-refused) and [ERR_CONNECTION_TIMED_OUT](/guides/err-connection-timed-out).

When a scripted request or deployment command fails

Capture the exact action, requested hostname, timestamp, response status, and error body. Compare the same operation through the dashboard or browser when possible. This helps distinguish authentication or project-configuration problems from build failures, transient network errors, and broader platform incidents. Do not expose deployment credentials or tokens in diagnostic output.