Why not use the client to send a disable message for the duration of your down window as part of your EOD process? That way it would be in your local client time zones. Send the length of your down window in the disable message and then the server will resume monitoring at the expiration of the down window, or it can send an enable message when the EOD process is complete.
Thanks,
Scot Kreienkamp
From: Brand, Thomas R. [mailto:TRBrand at cvs.com] Sent: Wednesday, July 08, 2009 3:34 PM To: hobbit at hswn.dk Subject: [hobbit] Handling Downtime across timezones?
I have client systems spread over several time zones from AST (Atlantic Standard Time) to HST (Hawaii Standard Time) and everything in between.
My hobbit server is in EST time zone.
The client systems run an 'End of Day' [EOD] process every night.
This process starts sometime between 12:30 am and 3:30am client LOCAL time and runs about 15 minutes.
During the EOD run, several background processes are intentionally stopped.
I don't want to have alerts generated if these processes are down/stopped during EOD but do want to be alerted if the processes are down outside of the EOD window.
E.g.:
client systems in Puerto Rico run EOD at 12:30am AST, which is 11:30 pm EST (previous day)
client systems in Hawaii run EOD at 3:30 am HST, which is 9:30 am EST
I have not figured out a way to do this - at least not with the DOWNTIME flag.
It appears that the DOWNTIME flag on the hobbit server uses the hobbit servers' current time to determine if an event is occurring during the DOWNTIME interval; that's a problem for me since from the hobbit servers perspective, the EOD can happen over a 10 hour period.
Any suggestions on how to configure things so that I don't get alerts for processes being 'down' from 12:30 - 3:30 am client local time?
N.B - I'm going to be dealing with over 7000 clients
Thanks,
Tom Brand