CPU-Basic Spaces hitting quota limits despite “unlimited hosting” — inconsistent behavior with Restart and Factory Rebuild

Spaces
1

I’m encountering inconsistent and possibly contradictory behavior regarding quota limits on Docker Spaces using the free-tier CPU-Basic hardware.

Summary of the issue

A Space that was previously running normally entered sleep (auto-sleep). When attempting to restart it, I encountered quota-related errors that prevent it from running again.

Observed behavior

  1. After auto-sleep, clicking “Restart Space” results in:

“You’ve reached your null quota limit, please upgrade your account, or pause your previous Spaces to restart this one”
→ Redirects to a 403 page

  1. As a workaround, I previously used “Factory Rebuild”, which successfully restarted the Space.
  2. However, since ~2 days ago, even Factory Rebuild now results in:

“Space Paused — You’ve reached your hardware quota limit across all your Spaces!”

  1. This occurs even when:
  • I have very few or zero active running Spaces
  • I tested with a completely new account (no Spaces at all), and the same error occurs

Expected behavior

According to the documentation, CPU-Basic Spaces are supposed to allow unlimited hosting (with sleep behavior). Therefore:

  • Restarting a sleeping Space should not hit a quota limit if no hardware is actively in use
  • Factory Rebuild should not immediately trigger a global hardware quota error
  • A fresh account with zero usage should not hit any quota

Questions

  • Is there a recent change in quota policy for CPU-Basic Spaces?
  • What does “null quota limit” refer to in this context?
  • Are sleeping Spaces still counted toward hardware quota?
  • Why would a fresh account hit a quota limit immediately?

Additional notes

  • This behavior started very recently (around 2 days ago)
  • The issue affects both Restart and Factory Rebuild flows
  • The error messages seem inconsistent (“null quota” vs “hardware quota”)

If this is an intended change, clarification in the documentation would be helpful. Otherwise, this may indicate a bug in quota tracking or Space state management.

2

Hi @rikunarita It looks like you’ve reached the current Spaces quota. Upgrading to PRO will help unlock additional quota for your account.

Plus, you’ll get access to PRO perks like:

  • $2 in free inference credits
  • 25 minutes of ZeroGPU daily
  • increased repo storage
  • …and more :fire:
3

Thanks for your response.

I’d like to clarify something regarding the CPU-Basic Spaces.

According to the documentation, CPU-Basic Spaces are described as being freely available for hosting (often interpreted as “unlimited” in terms of usage, aside from sleeping behavior). However, the current behavior seems to contradict that expectation.

In my case:

  • I cannot restart a sleeping Space due to a quota error
  • “Factory Rebuild” also fails with a hardware quota limit
  • The same issue occurs even on a fresh account with no active Spaces

This raises an important question:

Are CPU-Basic Spaces actually subject to a strict hardware quota limit?

If so:

  • What are the exact limits (number of concurrent Spaces, CPU allocation, etc.)?
  • Are sleeping Spaces still counted toward this quota?
  • Why would a completely new account immediately hit this limit?

Also, the error messages are inconsistent:

  • “null quota limit”
  • “hardware quota limit across all your Spaces”

This makes it difficult to determine whether this is expected behavior or a bug.

Could you clarify:

  1. Whether CPU-Basic is truly “unlimited” in practice
  2. What quotas are applied behind the scenes
  3. Whether this behavior is intentional or a recent change

This clarification would help avoid confusion for developers relying on CPU-Basic Spaces.