Our cloud provider's capacity in the us-central1 region hasn't recovered yet, and some sandboxes and labs with virtual machines still take a minute or two longer than usual to start during peak hours. We're rolling out further improvements to shorten these delays, and we'll post an update when anything changes.

Our cloud provider is still short on compute capacity in its US central region. With the improvements we rolled out last week, sandboxes and labs in our US region start successfully, but during peak hours (around 13:00 to 17:00 UTC) some sandboxes and labs with virtual machines take a minute or two longer than usual to start. Building custom virtual machine images in the US region may still fail occasionally.

We continue to work with our cloud provider and monitor the situation, and we'll post an update when anything changes. If you use hot start pools in the US region, we still recommend creating them well ahead of your event.

We've rolled out improvements that let sandbox creation work around the capacity shortage at our cloud provider's US central region. Since then, sandboxes and labs in the US region are starting reliably again. During peak hours (around 13:00 to 17:00 UTC), some sandboxes and labs with virtual machines may still take a minute or two longer than usual to start. Building custom virtual machine images in the US region may still fail occasionally.

Our cloud provider has not fully restored capacity yet, and we continue to monitor the situation. If you use hot start pools in the US region, we still recommend creating them well ahead of your event.

Our cloud provider is currently short on compute capacity in its us-central1 region. This affects sandboxes and labs hosted in our US region, mostly during peak hours (around 13:00 to 17:00 UTC).

What you may notice:
- Sandboxes and labs with containers or virtual machines take several minutes longer than usual to start.
- Hot start pools keep retrying until their sandboxes are created, so they may take longer to fill.
- Sandboxes and labs started on demand (not from a hot start pool) may fail to start in a small number of cases.
- Building custom virtual machine images in the US region may fail.

Sandboxes and labs in our European and Asian regions are not affected.

What we're doing:
- We're working with our cloud provider to restore capacity.
- We're rolling out improvements to make sandbox creation more resilient to capacity shortages.

What you can do:
- If a sandbox or lab fails to start, please try starting it again. Most retries succeed within a few minutes.
- Create hot start pools in our European or Asian regions instead of the US region.
- If you keep hot start pools in the US region, create them well ahead of your event so they have time to fill.

We'll post an update as the situation changes.