The router is happier, Cisco support is helping.
Ironically, it seems that netflow (a protocol for traffic analysis, which we’ve been using to help figure out what might be going wrong) might be part of the problem.
And I think I mentioned this in a comment on another article, but amidst all the other stuff that went wonky after the power was restored, the router that’s being problematic now lost its active supervisor card (which is basically a router’s CPU) early yesterday. The hot spare supervisor kicked in just like it should have, but just another fun little heartstopper!
More word when we have it!