Just run into this one. I've heard and read about default route or even static routes being affected by the routes with ping enabled, but not the connected routes.
It was also very difficult to pinpoint: there were to static routes via VTIs to AWS with ping enabled.
Disabling the ping on those and rebooting the unit did not revive it.
Removing all static routes except default and the one allowing communication with the management server, which was in another network, solved the issues.
Adding the routes back one by one did not result in the same behavior.
Pretty annoying, especially in remote upgrade situation. Thankfully there was someone there who could get me a console access to the unit.
routed was running. sk92511 not applicable. Behavior was similar to one described in sk107185 but more severe and it was on R80.20, which was supposed to fix it. Perhaps JHFA T_33 does address this, but in this case it should really be baked into the base image.