Hobby project stuck on "Resource provisioning failed" — legacy backend settings

Help
1

Project: claudemadedashboard1 (prj_YASvLKmTDyGlLCorLgrbzsKeCGy4) · Hobby plan
Since 2026-05-30 02:12 UTC, every deploy fails identically: Resource provisioning failed (BUILD_FAILED), 0 ms duration, build container never starts. The day before, 12 consecutive deploys succeeded. Nothing in my code changed between them.
A support reply identified the cause: my project’s backend has premium-only settings stuck on it that Hobby can’t provision —

  • isUsingActiveCPU: true
  • functionTimeout: 300s (Hobby cap = 10s)
  • SKIP_NAMESPACE_QUEUE
    My code is clean. vercel.json is just an SPA rewrite; repo-wide grep for those keys returns zero matches.
    Already cleared via the public PATCH API:
  • resourceConfig.functionDefaultTimeout: 300 → 10 :white_check_mark:
  • resourceConfig.fluid: true → false :white_check_mark:
    Result: still fails. The two remaining flagged settings (isUsingActiveCPU, SKIP_NAMESPACE_QUEUE) aren’t exposed in the public project API.
    Ask: please clear those two stuck backend settings on prj_YASvLKmTDyGlLCorLgrbzsKeCGy4. Last successful deploy was dpl_7CcDojyexEZgtrZHNuBD7uVwtK1o (2026-05-29 07:14 UTC) for reference.
    (Side note for the community: the Resource provisioning failed error doesn’t hint that the cause is plan-incompatible legacy config — I burned several hours assuming transient infra before support clarified. Worth documenting if this is a recurring post-downgrade pattern.)
2

problem is resolved by remote supabase integration, use API or token instead.