503 Service Unavailable

The server is temporarily unable to handle the request due to overload or scheduled maintenance. The condition is expected to be resolved. Retry later.

What It Means

503 Service Unavailable signals that the server is currently unable to process requests, but the unavailability is expected to be temporary. Unlike 500 (which indicates an unexpected fault), a 503 is often issued deliberately — the server knows it cannot serve traffic right now and is telling clients to back off.

The server may include a Retry-After header to indicate when the service is expected to recover. The value is a number of seconds to wait, or an HTTP-date after which the client may retry.

HTTP/1.1 503 Service Unavailable
Retry-After: 120

When It Occurs

• Scheduled maintenance. The operator takes the service offline intentionally and returns 503 so clients and search engines know to retry later rather than conclude the site is gone.
• Overload / traffic spike. The server has hit capacity — too many concurrent requests — and is shedding load by refusing new ones until the queue drains.
• Origin down behind a CDN or load balancer. The CDN or load balancer cannot reach any healthy origin instance and returns 503 to the client. This is distinct from 502, where the upstream returned an invalid response; a 503 means the service is temporarily unavailable due to overload or maintenance (per RFC 9110), rather than the upstream returning a malformed reply.
• Health-check failures. A load balancer removes all origin pods from rotation due to failed health checks, leaving no backend to forward requests to.

How to Diagnose

Use the Probe tool to check the status code and any Retry-After header. Check the cloud provider status pages (see Cloud Status) to determine whether an upstream incident is causing the outage. If you operate the server, check load metrics, health-check results, and application logs to determine whether the cause is capacity, a failing dependency, or a deployment issue.

Distinguish from related codes: 500 is an unexpected application crash; 502 is a proxy receiving a bad response from an upstream; 503 is the server consciously refusing requests due to temporary unavailability.