That 'broken' third hole from the top ... shows probability that the bolting immediately above it lost 'clamping' torque and all the load began to bear on the 'projected semi-circumferential area (top side of the third bolt hole - acting like a 'saddle'), probably (as shown) began to develop 'lateral cracks' AT between the top half and bottom half of the bolthole ... the crack began to flex which accelerated FATIGUE failure.

You need to properly torque such bolting connection and with 'sufficiently large' and sufficiently THICK washers under the 'nuts' so that the loading is transferred by the 'clamping force' to the mating parts together due to proper bolt torque .... not the 'strength of the bolt' in shear.

Several ways to prevent this in future: (much) THICKER chainplates; proper torquing (and regular routine maintenance of the torque) using large and thick 'washers' under the 'nuts' or bolt head.

When you 'hang' a load on a bolt in shear (always a BAD idea), instead of making this a 'friction' joint which 'distributes the loading' by properly torquing, the top of the 'hole' becomes whats known as a 'stress riser' ... especially at the circumference of the bolt hole that is at 90° from the 'pull'.
In a friction joint such as this, the only contribution of the bolts should be only to 'drive the mating surfaces together to create the friction', and only act as a 'backup' if the torque on the bolts is lost.

Just my 'engineering' opinion.

Rx:
Thicker Plates
Use a torque wrench, and on regular maintenance basis.