You said you used 3M bubbles (micro-balloons maybe?) for the thickener... My CLC instruction manual says to use Silica powder (Cab-o-Sil) in the scarf joints. I guess I don't know the difference between your 3M bubbles and Cab-o-Sil except I envision the bubbles to be more for filling gaps as opposed to connect wood. Hopefully people here can educate me.

The Cab-o-Sil I used in the scarf joints made a very strong joint. In fact, I made up a test piece and the wood failed before the joint (somewhat similar to yellow wood glue joints). If the joint was over compressed before the epoxy hardened, the joint could have been "starved", resulting in a weak joint. Did you allow the epoxy to harden a sufficient amount of time prior to flexing the wood? When I did my scarf joints, I did them in a 55 - 60 F cellar and I let them sit 48 hours prior to moving them.

If the joints are weak, it's best to make them all fail at the wood interface and start again. I don't know why you can't remove the epoxy from the scarfs (probably with a belt sander - don't forget the dust mask) and re-epoxy provided you can remove the weak epoxy and get down to wood again. This method will allow you to keep roughly the same boat length. Hopefully other people here will contribute as well and give other suggestions.