Skip to main content
https://img.shields.io/badge/issue_tracking-github-blue?logo=github Check the qBraid Status Page to see if there are any ongoing incidents or maintenance windows affecting services. If you encounter an error that isn’t included on this page or if the solution provided doesn’t work, please submit a support request, email us, or open a bug report. Also see our FAQs.

​
Launch failed: Compute hours exhausted

When you click Launch on a subscription (CPU) profile and get:
you have used all of the compute hours included with your plan for the current month, and you do not have credits set up to cover the session instead. The amounts shown in the message vary with the profile you tried to launch. Nothing is wrong with your account, and none of your files or environments are affected. Compute hours reset at the start of each month. Until then you can still download your files, and you can restore the ability to launch by upgrading your plan or by switching sessions over to credits.

​
Get your files without launching

If you only need your files back, you do not have to launch Lab at all — download them from the Usage Overview section of your account dashboard. See Download your files.

​
Upgrade your subscription

A higher tier comes with a larger monthly compute hour allocation (Free 100, Standard 400, Pro 2000). See Personal Subscriptions to change your plan, or Organization Subscriptions if your seat is managed by an organization.

​
Buy credits and enable auto-switch

Credits cover CPU sessions once your compute hours run out, but only when auto-switch is turned on. Both steps are required — adding credits on its own will not resolve the error.
  1. Buy credits from Account > Wallet. If you are in an organization without purchase permissions, request credits from an admin instead.
  2. Enable Auto-switch to credits under Account > Usage > Compute Settings.
Sessions then bill against your credit balance at the per-minute rates listed in Pricing. You can confirm your remaining compute hours and credit balance at any time on the Usage page.
On-demand instances (GPUs and high-spec CPUs) always bill credits directly and never draw on compute hours, so they remain available even when your monthly allocation is exhausted.

​
Launch failed: JupyterHub is temporarily unavailable

When you click Launch and get an error like:
this is a temporary problem on qBraid’s side, not something wrong with your account. It is unrelated to your quota, credits, subscription, or anything you did — your files, environments, and balances are all unaffected, and no compute hours are consumed by the failed launch. The user, cluster, and operation named in the message are diagnostic details, not settings you need to change. The service often recovers on its own, so it is worth waiting a few minutes and trying again:
  1. Check the qBraid Status Page for an ongoing incident.
  2. Wait a few minutes, then click Launch again.
  3. If it still fails, submit a support request from your account page. Include the full error message — the user, cluster, and operation it names help us pinpoint the problem.
  4. Once we confirm the service is restored, launch again as normal.
Repeatedly clicking Launch will not speed up recovery. Give it a few minutes between attempts.

​
Lab server errors

If you get one of the following errors from inside Lab,
  • Service unavailable or unreachable
  • File Save Error
For these types of errors, refreshing the page won’t work and you will need to restart your qBraid Lab server. You can stop your running server from your account dashboard, and then launch a new instance.

​
ModuleNotFoundError

While running a notebook, if you get a ModuleNotFoundError after an import statement:
  1. Check to make sure you are using the correct notebook kernel for your environment, see Switch notebook kernel.
  2. If you are using the correct kernel, the package you are trying to import may not be installed in that environment. See Install new package.

​
No space left on device

If running a notebook cell produces errors like:
you have run out of disk space. The exact wording varies depending on which library tried to write to disk first, but the giveaways are database or disk is full, [Errno 28] No space left on device, and complaints about cache directories that cannot be written. There is nothing wrong with your code — there is simply nowhere left to save anything. Check your current usage against your quota from the Storage card in the Settings extension, opened with the Settings icon in the bottom-left of Lab, or from the Storage card on the Usage page under Account > Usage. Then either free up space or raise your quota:
  • Free up space. Installed environments are usually the biggest consumers, so uninstalling ones you no longer use frees the most space. You can also delete files from the file browser or track down large directories from the terminal. See Disk Usage Management.
  • Upgrade your subscription. Storage scales with your plan — 15 GB on Free, 50 GB on Standard, 150 GB on Pro. See Personal Subscriptions.
After clearing space, restart the kernel (Kernel > Restart Kernel) so the affected libraries can recreate their caches.
If you stay over your disk quota, your Lab session will eventually be terminated, and you will need to reduce your usage before you can resume. See Exceeding Disk Usage Quotas.