As ER333 has pointed out, the undermining due to the flawed drainage system design produced (a) thinning leading to diagonal cracking that keeps getting worse, (b) voids under the slab due to wash-out which now hold more water for more washout, etc etc [except where voids are filled in recent repairs] and (c) the spillway walls were not ripped out previously and now the outflow has potential damage every time it is ramped up. (d) I think we are also seeing a further compromised underslab drainage system itself, water release may occur without the full release of pressure.
So maybe, due to the flawed drainage system design, the one-two punch of initial weakening in earlier, higher outflows, and then later, final breaking at much lower outflows.
A 1967 postcard image reveals a natural "Water Percolation Seam" at/near the Blowout Failure location. (note: the postcard was digitally enhanced to filter the "texture dots" to provide this clearer forensic image).
As this photograph was taken prior to the pour of the spillway concrete slabs, the image reveals the "grade" construction of which the slabs were to be emplaced upon. This forensic "seam" presence provides geologic insight, especially regarding the question of "Where's the Missing Water?" from the spillway drains in this pre-blowout failure location. The photograph also provides insight into how "un-captured" waterflow in the sub-par drain design could "wash" a void layer between the bedrock and the concrete slab pour. The spillway "chute" design has embankment fill. There were no signs of embankment erosion near this failure area. The only sign was a preference of large tree growth (roots + growth = subsurface water). The "Seam" answers the question of a "deeper" subsurface flow capability. Over time, the erosion of this flow could increase the underslab voiding.
Given these conditions, the deep waterflow + erosion could have created substantial voids (note: persistent spillway "missing water" at equal levels of the "working drains" observed flow rates = a high capacity deep flow "seam" likely was created in time; exploiting this location's natural formation). Each time the spillway was operated, it eroded and enlarged a void area. Until a higher structural stress condition was applied - such as a higher flow rate - did the right conditions of the voiding stress the failed slab to its limits. Part of this equation is the other dynamic stress condition of hydraulic jacking (one is up the other is down).
Subsurface Seam at Blowout Failure location. Photograph provides insight into how "un-captured" waterflow in the sub-par drain design could "wash" a void layer between the bedrock and the concrete slab pour.