3 minTroubleshooting

Instance Failed to Start

Common causes and fixes for deployment failures.

1

Check available provider capacity

GPU availability varies by provider. If a deployment fails immediately, the selected GPU may be out of stock. Try switching providers in Deploy → Smart Deploy or choose a different GPU tier.

2

Review instance logs

Navigate to Dashboard → Instances → Logs and look for startup errors. Common errors:

  • CUDA version mismatch — Select a different base image in your preset
  • OOM during init — Reduce the number of containers or choose a higher VRAM GPU
  • Image pull failed — Check the Docker image name and registry credentials
3

Verify your credit balance

Deployments are blocked when your credit balance reaches zero. Visit Dashboard → Billing to top up credits.

Still failing?

If none of the above resolve the issue, contact our team directly — include the instance ID and any log output.

Back to Docs