Unfortunately, I just stopped the rule, which removed all the scheduled jobs.
I also re-wrote it to simplify the logic to prevent multiple runs using the technique you suggested on this post.
Next time it happens I will try hitting "update" to see what happens, but even if hitting update cleans out the backlog of scheduled events, it won't solve the problem. This has been going on for several months -- it just took me a while to narrow it down to this rule.
