It's something I'll consider. If it's a big annoyance, one current workaround is hiding that worker. Hidden workers will not generate idle notification emails. Idle miner code is something I have been meaning to rewrite the script for to find a more efficient way to scan through the last share times of workers, so once I get to that (right now my focus is DDoS mitigation + new colocation), I will likely fit in the per-worker notification settings.
I would like to see code for a minimum threshold too. One of my ASICs dropped off and required a restart to get it active again.... The others kept churning away and I did not notice it until I saw the green light was on as I walked by... Not sure if that is really possible though due to variance over time..
Unfortunately, anything related to a worker's actual speed will need to be a local warning system. It's much more reliable, and more feasible to implement. Some kind of API monitor that sends you an email when a worker drops too low would not be hard to implement, whereas offering it as a standard feature could add a lot of extra load to the pool given how many actual workers are running at any given time.