Skip to main content
HTTP status: 503 · Title: “Service unavailable” · Retry-After: computed per response

​
When it fires

A service the endpoint depends on (for example ZeroID’s database or a blockchain RPC provider) could not be reached, and the endpoint cannot answer without it. Almost every endpoint can return it. It is distinct from INTERNAL_ERROR (“something unexpected happened”) and from ROUTE_DISABLED (“ZeroID disabled this on purpose”).

​
Why it happens

  • A transient infrastructure blip (most common; clears in seconds).
  • A sustained outage of a dependency that ZeroID is already working on.

​
How to fix

  • Honour Retry-After and back off with jitter. Most occurrences clear within one retry.
  • The TypeScript SDK raises this as KashServerError and retries it automatically within its retry budget.
  • Retry writes with the same Idempotency-Key so a retry cannot execute twice. See Idempotency.
  • If retries keep failing for more than a few minutes, contact ZeroID support with a requestId.