503 during traffic spikes?

Add caching, autoscaling rules, and protect origins with a queue or WAF.

Should search engines index 503 pages?

Usually no—use temporary 503 with Retry-After for real outages; avoid soft-404 content.

Check throttling quotas and burst limits in the gateway configuration.

Error fix

The server is temporarily unable to serve traffic—overload, maintenance, or dependency outage.

Start here first. Step 1 fixes most cases—then work down the list.

Why this works

These steps work because it clears stale cached state that can cause the same request to fail repeatedly.

503 tells clients the service exists but cannot handle the request right now. Retry-After may hint when to try again.

Capacity limits
Rate limits, connection saturation, or autoscaling lag under load.
Maintenance
Planned downtime or feature flags blocking requests.

Restore capacity
1. Check CPU/memory, queue depth, and error rates in your platform dashboard.
2. Scale out or shed load (queue, cache, circuit breakers).
3. Return Retry-After when appropriate so clients back off.

Also see: 503 — browse the HTTP status hub.

503 during traffic spikes?: Add caching, autoscaling rules, and protect origins with a queue or WAF.
Should search engines index 503 pages?: Usually no—use temporary 503 with Retry-After for real outages; avoid soft-404 content.
503 from API gateway?: Check throttling quotas and burst limits in the gateway configuration.

Still stuck? Try these related fixes next.

Tweak the message and run again—we'll match an existing fix or generate a new page.

Original error message

503 Service Unavailable