There is a known bug in the update of the internal Workgroup/VGrid cache used to speed up various user access decisions instead of traversing the entire state/vgrid_home/ hierarchy each time. It results in the cache sometimes going out of sync when multiple users are updating participation lists under heavy site load so that client HTTP requests may time out before completing. In practice some of the participation updates may get lost in the cache and effectively remove the usual owner admin (wrench) links to owned Workgroups.

It should of course be fixed e.g. by further reworking the concurrent updates, or better, moving the clunky cache updates into a single daemon process to also rip out the lock congestion and client page load times associated with the current on-demand updates.

For the time being the owners can still fix the issue themselves by manually opening the admin page URL and use the Repair button at the bottom. For the Workgroup/VGrid called MYSTUFF the link would be:
https://migrid.site.url/wsgi-bin/adminvgrid.py?vgrid_name=MYSTUFF