I published a post on dev.to about debugging a mysterious total blocking time issue for ColorBliss: