It appears to be something that Databricks currently does not expose neither though UI or API.
My solution was to delete the entire metastore. Once metastore was deleted, the workspace in this metastore was unable to show the catalog page. It was blinking and broken.
After deleting metastore I deleted the workspace.
Deleted metastore still can be queried via API, but all workspaces on it will throw error on getting. They only show in the account.workspaces.list command as a list of IDs. This metastore data seems to be going to sit there forever.
I've created new workspace, after none left. New workspace sits in the different instance of the metastore, and so it does not have naming conflict anymore.
I would say this situation is very difficult to resolve, and I would have had a weirdly named default catalog forever if I already had other workspaces in the region.
Perhaps having own metastore storage is a better solution to it.