[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [hobbit] Delayed recovery (was: bbretest-net.sh)



Hi,
for me it's hard to say if it should be per host or per test or what.
I attach an example from my histlog to show what I had and what I wish.This is an extract from the conn-test:


Sun May 08 01:50:11 2005  green  4:59:18
Sun May 08 01:49:44 2005  red  0:00:27
Sun May 08 01:48:31 2005  green  0:01:13
Sun May 08 01:48:10 2005  red  0:00:21
Sun May 08 01:33:55 2005  green  0:14:15
Sun May 08 01:31:52 2005  red  0:02:03
Sun May 08 01:29:02 2005  green  0:02:50
Sun May 08 01:27:42 2005  red  0:01:20
Sun May 08 01:26:41 2005  green  0:01:01
Sun May 08 01:12:45 2005  red  0:13:56
Sun May 08 01:11:40 2005  green  0:01:05
Sun May 08 01:11:38 2005  red  0:00:02
Sun May 08 01:11:19 2005  green  0:00:19
Sun May 08 01:09:55 2005  red  0:01:24
Sun May 08 01:07:17 2005  green  0:02:38
Sun May 08 01:05:41 2005  red  0:01:36
........
........
since Thursday 06:20

This I want to appear as:

Sun May 08 01:50:11 2005  green  4:59:18
Sun May 08 01:48:10 2005  red  x:xx:xx
Sun May 08 01:33:55 2005  green  0:14:15
zzzzzzzzzzzzzzzzzzzzz      red           x:xx:xx

Regards
Lars
Hobbithobbyist

Http://leopg9.no-ip.org
----- Original Message ----- From: "Henrik Stoerner" <henrik (at) hswn.dk>
To: <hobbit (at) hswn.dk>
Sent: Saturday, May 07, 2005 6:16 PM
Subject: Re: [hobbit] Delayed recovery (was: bbretest-net.sh)



On Sat, May 07, 2005 at 07:02:48AM +0200, lars ebeling wrote:

I would like to have an option to bbretest-net so that I could configure how long time a service should be up before changing status.

I have since 2 days a very unstable internetconnection, that goes from red
to green to red........
Perhaps green for 30 secs.

I think this is a more general issue, not just for the network re-test.

Currently, if a status goes from red -> green, it clears all of the
alert- and acknowledge information. I think it would make sense to delay
that slightly - so that if it goes red again within that short period of
time, it will not restart all of the alerts and require a new
acknowledge for the alerts to go away.

How long should this delay be - should it be configurable on a per-host
or per-test basis ? Or just a global setting ?


Henrik


To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe (at) hswn.dk