If you belong to multiple workspaces, you may have noticed the switcher sometimes needed a few attempts before the workspace actually changed. The cause was a race between your session updating and in-flight requests quietly overwriting it — and for workspace owners, a stale session could stick around instead of correcting itself.
Both problems are fixed. Switching now verifies the new workspace is fully active before reloading, creating a new workspace waits until it's ready before opening the dashboard, and any stale session state now heals itself within a few minutes. If a switch genuinely can't complete, you'll see a clear error instead of a silent no-op. Nothing to do on your end.