Pilot draining: add the possibility of specifying the downtime lenght
In the current downtime script, a pilot that starts after the specified downtime will ignore the downtime: , https://github.com/glideinWMS/glideinwms/blob/master/creation/web_base/check_wn_drainstate.sh#L78-L79 ,
The logic is that if there is a downtime then pilots should not start, so it is probably a stale file. And in order to not waste CPUs the pilot just run normally.
A site admin using this feature was confused by this, and argued the downtime just started and pilots should have not accepted jobs. In order to be clear it would be best to specify the downtime lenght.