Feature request: Long TTL TCP connections and hard-stop-after

HAProxy can allow graceful termination and settings how to deal with the duration of “gracefulness”
specifically hard-stop-after

coming from the IoT world, where devices are keeping connections forever and reconnects on connections lost, I see a nice to have feature:
after waiting for the connection drain, hard-stop-after, be able to configure how remaining connections are terminated, either as a rate: connections/sec or simple duration, that will allow to add an additional degree of gracefulness and avoid reconnect storm.

as example:
hard-stop-after 300s
terminate-after 60s

so it will wait 300 sec to drain and will terminate remaining connections during the next 60s window, evenly.

Avoiding reconnect storm makes sense indeed.

Can you file a feature request on github?