intel · Rate Limits
Intel Rate Limits
Intel does not publish per-second rate-limit numbers in the public Intel Trust Authority REST API documentation. Limits are operationally enforced as fair-use throttling on shared infrastructure. Intel oneAPI portals are documentation / download surfaces rather than runtime APIs and do not have per-call rate-limit semantics worth capturing here.
2 Limits
Throttle: 429
Confidential ComputingAttestationoneAPIRate Limiting
Limits
Intel Trust Authority attestation throughput api-key
fair-use throttling on shared attestation infrastructure
Regional availability region
US/World endpoint and EU endpoint; unavailable in mainland China
Policies
Backoff Strategy
Clients receiving 429 or 503 should implement exponential backoff with jitter and honor Retry-After when present.
Region Selection
Use the EU endpoint (https://api.eu.trustauthority.intel.com) for EU-resident workloads and the global endpoint (https://api.trustauthority.intel.com) elsewhere; the service is unavailable in mainland China.
Token Caching
Cache issued attestation tokens for the duration of their validity rather than re-attesting on every workload start to reduce request volume.
Fair Use
Intel reserves the right to throttle traffic that materially impacts shared infrastructure; paid support customers can request guaranteed-throughput SLAs.