Good practice would detail as Mike suggests, with DBAs providing the tie forces. This very likely has these, otherwise I would expect the crack to result in detachment of the embed. There may also have been enough restraint to cause this crack, depending upon how the structure above was configured. I have frequently seen construction where the embed would have been better if it were farther into the recess, but it was marginal as constructed (not truly deficient.)
The crack initiated under shear force as you surmise, with or without restraint of the embed by structure above. The crack stopped moving once the reinforcement was mobilized. If the reinforcement/DBAs were stressed significantly, but within yield, the crack would widen under load and then squeeze back together when unloaded.
I agree that the design of the support, and the size and placement of the embed may have been inadequate for the forces involved. But, if the reinforcement is adequate and there is sufficient shear friction developed by aggregate interlock, this member may still be serviceable. Is there any vertical displacement? And do you have shops for the embed?