Headless storefront API down — how to detect and fix it
Quick answer
On a headless store the frontend and the commerce/API layer are separate, so the site can look perfectly fine while product, cart, or checkout API calls fail — an outage a homepage ping never catches. Causes include an expired API token, a rate-limited or changed backend API, a failed deploy of the API layer, or a CORS/config break. Detect it by probing the actual API responses, not just the page, and roll back the API deploy on managed infra.
Symptoms
- The storefront renders but products, cart, or checkout don't load data
- "Something went wrong" or infinite spinners on dynamic sections
- A homepage uptime check reports healthy while orders are impossible
- It broke right after an API-layer deploy, token rotation, or version bump
Common causes
- An expired or rotated storefront API token/key
- A backend/commerce API rate limit, outage, or breaking version change
- A failed or partial deploy of the API/BFF layer
- A CORS, environment-variable, or endpoint-URL misconfiguration
How to check
- 1.Probe the actual product/cart/checkout API endpoints, not just the homepage
- 2.Check the API layer's logs and recent deploys for errors
- 3.Confirm the storefront API token is valid and not rate-limited
- 4.Verify environment variables and endpoint URLs match the current backend
How to fix it
1. Test the API directly
Call the product, cart, and checkout endpoints your frontend uses. A healthy homepage means nothing if these fail — this is where headless outages hide.
2. Fix the token or config
Rotate/restore a valid API token and confirm endpoint URLs, CORS, and environment variables match the live backend.
3. Roll back the API deploy
If a deploy of the API/BFF layer broke it, revert to the last healthy version and verify a real product/checkout call succeeds.
4. Monitor the money path, not the homepage
Liulum probes your real revenue journey — including API responses — so a headless failure the homepage hides is caught in seconds, and on managed infra a bad deploy can be rolled back automatically.
How Liulum helps
Liulum watches this path continuously, tells you in plain English the moment it breaks, and on sites it manages can run a safe, allow-listed, reversible repair automatically — so an outage lasts seconds, not hours.
Protect your storeFrequently asked questions
Why does my headless store look fine but nothing works?
In a headless setup the frontend is decoupled from the commerce API, so the page can render from cache or static content while the API calls behind product, cart, and checkout fail. A homepage ping can't see it — you have to probe the API responses themselves.
How do I monitor a headless storefront properly?
Check the actual API/GraphQL endpoints and the full revenue journey (product → cart → checkout), not just page load. That's the only way to catch an API-layer outage before customers do.
Can Liulum monitor and repair a headless stack?
Yes — Liulum monitors your real revenue journey including API responses, and on managed infrastructure can roll back a breaking API deploy automatically, verified and reversible, while alerting you with the plain-English cause.