Anam 99.9% uptime SLA vs Sume failed-job retry metadata and refunds

Sume does not publish an uptime percentage in the docs cited here. It documents what happens per job: failed jobs expose category, stage, retryability and next action, and the reservation is released or refunded. Anam states a 99.9% uptime SLA, SOC 2 compliance, and connection times.

Anam facts are from its Cara-4 post (updated Sept 1, 2026); Sume facts from Errors and credits and Generation admission, read 2026-10-01.

What does Anam state?

For recent traffic, p50 connection time is under one second and p95 under three. Anam says Cara-4 is backed by a 99.9% uptime SLA and SOC 2 compliance.

What does a failed Sume job tell you?

Failed jobs expose public error metadata such as category, stage, retryability, retry-after seconds, public reason and next action. Internal provider payloads are not public fields. Failed jobs and failed queue admission release or refund the reservation where applicable.

How should my code handle failure?

Submit async with an idempotency key, as the admission docs advise for production. Sume sends terminal job events only, so handle job.failed and read the error object; only retry when retryability says so. See Sume webhook retry schedule.