Hi Alex,
as the 6.50 is a beta version, I place this post in the installation and configuration forum (instead of the wish list). I hope this is ok.
In several posts in this forum the "warning" status was desired to signal a condition which is less critical than the "bad" status".
Therefore many users and I do appreciate very much your work on this status.
The warning status obviously can be used in two ways:
a) introducing a second threshold for the test, which is lower than the threshold set in the test dialogue
b) introducing a recurrencies threshold.
In both cases the expressions required could be simplified:
Suggestion:
It could be useful in case of
a) a tick box like "check condition only if test condition is fulfilled".
This would simplify the expression used: Only one threshold necessary instead of two.
E.g. Disk space test. Limit 1 GB
Without the check box you always have to declare two limits in the expression like ('%Reply%'<='3 GB') and ('%Reply%'> '1 GB'). The limit '1 GB' is redundant to the settings in the test properties.
b) A variable like %RecurrenciesOfSimpleStatusIsNotOK%.
This would simplify the expression as already explained in the thread "Normal Status and %Recurrencies%"
Kind regards
Thomas
WARNING Status
-
- Posts: 166
- Joined: Sat Apr 15, 2006 2:14 pm
- Location: Germany, Bavaria
Sounds like "check condition only if test condition is fulfilled" is equivalent to '%SuggestedSimpleStatus'=='DOWN'. Doesn't it?a) a tick box like "check condition only if test condition is fulfilled".
What exactly this variable should represent?b) A variable like %RecurrenciesOfSimpleStatusIsNotOK%.
Regards
Alex
There are two ways (always ?) to perform tests and that leads to different use of expressions
1st
The test is setup as:
Alert when CPU Usage (Or Bandwidth, ..) usage is more than 60%
Set NORMAL status if
("%SuggestedStatus%"=="Ok") and ("%SuggestedReply%">"30 %") and ("%SuggestedReply%"<"50 %")
Set WARNING status if
("%SuggestedStatus%"=="Ok") and ("%SuggestedReply%">"49 %")
2nd
The test is setup as:
Alert when CPU Usage (Or Bandwidth, ..) usage is more than 30%
Set NORMAL status if
("%SuggestedStatus%"=="Bad") and ("%SuggestedReply%">"30 %") and ("%SuggestedReply%"<"50 %")
Set WARNING status if
("%SuggestedStatus%"=="Bad") and ("%SuggestedReply%">"49 %")
1st
The test is setup as:
Alert when CPU Usage (Or Bandwidth, ..) usage is more than 60%
Set NORMAL status if
("%SuggestedStatus%"=="Ok") and ("%SuggestedReply%">"30 %") and ("%SuggestedReply%"<"50 %")
Set WARNING status if
("%SuggestedStatus%"=="Ok") and ("%SuggestedReply%">"49 %")
2nd
The test is setup as:
Alert when CPU Usage (Or Bandwidth, ..) usage is more than 30%
Set NORMAL status if
("%SuggestedStatus%"=="Bad") and ("%SuggestedReply%">"30 %") and ("%SuggestedReply%"<"50 %")
Set WARNING status if
("%SuggestedStatus%"=="Bad") and ("%SuggestedReply%">"49 %")
@Thomas
Sorry I have read that post to late.
Unfortunately I'll have to do some work during the day
I'll try tomorrow. You can delete your phone number
One of the things I'll do is to make tests more tolerant against "False-positives".
- Make a test go to BAD after 2 or 3 fails, not after the first one.
Not for all tests of course, but I have some candidates ...
Allow bandwidth and CPU usage tests to go over limits for some time.
(CPU and bandwidth peaks are absolutely normal)
Some kind of: OK -> Normal -> Normal -> Normal -> Warning -> Warning -> Bad
Drivespace and CountFiles tests go to Warning before Bad.
- That can be done by free Space (1-3 GB Warning, <1 GB Bad)
- And/Or by number of recurrencies
If ISDN kicks in (The ISDN Interface is up - tested via SNMP) I'll set those test to Warning, because the location can work, only the performance maybe degraded.
The same for Interface Errors.
If PING response times slow down for some time I'll call that a Warning
The main thing should be:
If there is a test RED someone has to look after that.
Today we often have RED tests but not really a problem, therefore sometimes real problems are ignored today.
Juergen
Sorry I have read that post to late.
Unfortunately I'll have to do some work during the day

I'll try tomorrow. You can delete your phone number
One of the things I'll do is to make tests more tolerant against "False-positives".
- Make a test go to BAD after 2 or 3 fails, not after the first one.
Not for all tests of course, but I have some candidates ...
Allow bandwidth and CPU usage tests to go over limits for some time.
(CPU and bandwidth peaks are absolutely normal)
Some kind of: OK -> Normal -> Normal -> Normal -> Warning -> Warning -> Bad
Drivespace and CountFiles tests go to Warning before Bad.
- That can be done by free Space (1-3 GB Warning, <1 GB Bad)
- And/Or by number of recurrencies
We use WAN routers with ISDN Dial-Backup.KS-Soft wrote:2) Probably %WarningStatusRecurrencies% and %NormalStatusRecurrencies% counters is good idea![]()
If ISDN kicks in (The ISDN Interface is up - tested via SNMP) I'll set those test to Warning, because the location can work, only the performance maybe degraded.
The same for Interface Errors.
If PING response times slow down for some time I'll call that a Warning
The main thing should be:
If there is a test RED someone has to look after that.
Today we often have RED tests but not really a problem, therefore sometimes real problems are ignored today.
Juergen