GitBook SaaS sync timeout handling for repositories exceeding 10‑minute build limit
0 reputation · 09 Mar 2023, 22:49 UTC
GitBook SaaS enforces a 10‑minute hard timeout on each Git sync operation, and repositories that exceed this limit fail silently with a generic 'build failed' status in the dashboard.
The legacy gitbook‑cli allowed unlimited local builds, but the modern platform runs builds in ephemeral containers with no user‑configurable timeout override, and partial sync failures — where some pages succeed and others time out — leave the space in an indeterminate state while the UI still reports 'Synced'.
The public API (`GET /spaces/{id}/sync`) returns `status: 'success'` even when individual page builds timed out, with no per‑file error array or `partial_failure` flag, and rapid re‑sync attempts hit an undocumented rate limit without a `Retry-After` header.
How will GitBook expose per‑file build errors in the sync response?
Is there any documented way to adjust the 10‑minute timeout for large repositories?
Will future API versions include a partial_failure indicator?
gitbook