Check service health and recover from blocked states

Check public service health#

1. Open the Health page

   Use the public language-specific Health route even when the authenticated workspace cannot load.

2. Read the overall state

   Healthy, degraded, unhealthy, and unknown summarize current checks; do not interpret unknown as confirmed failure.

3. Review the available checks

   Use the high-level status shown by the page to distinguish a broad service issue from a resource-specific problem.

4. Refresh once

   Use the manual refresh control when the displayed state may be stale.

5. Retry after recovery

   Return to the failed task only after the page reports a usable state.

Respond to Account Locked#

• Read the lock explanation and use the provided support path.
• Contact an organization owner to confirm whether access was intentionally disabled.
• Do not create a second account to bypass the organization decision.
• If other workspace memberships remain active, sign in and choose only an organization you are authorized to use.

Respond to Desktop Required#

• Authenticated organization routes require a browser width greater than 768 pixels.
• Widen the window, rotate only if the resulting width is sufficient, or use a desktop/laptop device.
• This state protects complex annotation, visualization, and training interfaces from unsafe small-screen operation.
• The documentation portal itself remains responsive, but the product workspace intentionally requires a wider viewport.

Recover from missing, loading, or request errors#

1. Read the page state

   List and detail pages distinguish loading, empty, error, and not-found states. Wait for loading to finish before retrying.

2. Use Retry once

   For a retryable error, choose Retry and observe whether the message or status changes.

3. Use Back or Dashboard

   A missing resource or 404 page provides a safe route to a known location.

4. Confirm workspace and permission

   A valid resource can look missing when the wrong organization is active or access changed.

5. Collect safe evidence

   Record the resource type, non-sensitive visible status, approximate time, and exact safe error text.

Handle an application update prompt#

• Read the current run or upload state before refreshing.
• Save annotations and form changes first.
• Avoid refreshing during a local file transfer; server-side jobs can continue, but the page should be reopened and verified.
• After updating, confirm the active workspace and resource before continuing.