exp
This marks when the token should stop being accepted. The problem is that server time, proxy time, and the operator's local view can all differ.
🔒 Browser-side processing
A practical guide to reading JWT exp, nbf, and iat claims, including the incidents that come from clock skew and bad assumptions.
First Read
expThis marks when the token should stop being accepted. The problem is that server time, proxy time, and the operator's local view can all differ.
nbfThis marks when the token should start being accepted. Immediate failures right after issue time often point here first.
iatIt tracks issue time and is not always enforced directly, but it becomes useful when you compare all timing claims together.
Common Failures
JWTs usually use Unix seconds. When frontend code or logs treat them as milliseconds, the resulting timestamps look completely wrong.
Readable tokens are not automatically trustworthy. External issuer setups still need signature validation through JWKS or an equivalent check.
Even a small time difference between application and auth servers can create repeated boundary failures. Teams should document acceptable skew explicitly.
Workflow
exp, nbf, and iat into human time.