Time threshold

[summa_sp]

Like it

Up

0

Down

Drop it

::

Update: The problem only happens with modules sent from the agent.  A port check done by the server will honor the Time Threshold value, but a file system check that exceeds the value on the agent does not honor the Time Threshold value.  It sends an alert each time.

[daniels]

Like it

Up

0

Down

Drop it

::

What version of pandora are you using?

Check if the time interval is correct in the pandora_console in:

View agents -> Agent Detail -> Data -> int

The default value is 300 (in seconds). Set it to 60 and see if it works.

Regards.

[summa_sp]

Like it

Up

0

Down

Drop it

::

I’m using 3.0 RC2, and the time int is set to 60 seconds on the module, the agent, and the server.  I have done some more testing, and it seems to be an issue with modules configured in the agent.  A file system check sent by the agent causes an alert each time.  A network port check configured in the server works correctly. 

It’s easy to test, just do a file system check in the agent:
module_begin
module_name Disk Util /var
module_type generic_data
module_exec df -k /var | tail -1 | tr -d “%” | awk ‘{ print $5 }’
module_max 100
module_min 0
module_description /var disk util %
module_end

Let’s say the /var partition is 50% full. 
Configure the module to send Warning at 45%-60%.
Configure an alert template to send email when it receives a Warning alert.
Set the Time Threshold to 15 minutes, the Min and Max to 1, and the default action to send only 1 alert. (On 1)

The alert will email each time the alert fires (60 seconds), even though it is supposed to wait 15 minutes. 

[summa_sp]

Like it

Up

0

Down

Drop it

::

Further testing has revealed that it only seems to happen with the agent running on Solaris 10.  I have tested it with Window server 2003 and the OpenSUSE agent, and the Time Threshold works fine.  When I configure the identical alert from a Solaris 10 agent, it does not honor the Time Threshold value.

[summa_sp]

Like it

Up

0

Down

Drop it

::

The problem is that the alerts fires every other Interval for the agent update.  If the agent update interval is 3 minutes, I get an email ever 6 minutes, if the agent update interval is 5 minutes, I get an alert every 10 minutes, etc.  It seems like it’s not keep track of the alert state beyond one cycle.

[Sancho]

Like it

Up

0

Down

Drop it

::

Further testing has revealed that it only seems to happen with the agent running on Solaris 10.  I have tested it with Window server 2003 and the OpenSUSE agent, and the Time Threshold works fine.  When I configure the identical alert from a Solaris 10 agent, it does not honor the Time Threshold value.

This is probably because agent interval is set to 0 (not by you, possibly a bug), could you, plrase, post here the XML coming from this solaris agent ?.

Thanks

[summa_sp]

Like it

Up

0

Down

Drop it

::

I have narrowed the problem down to a bug with the Solaris agent, and a Time Threshold in the alert template (Step 2) of 5, 10, or 15 minutes.  If the threshold is 30 minutes or more, it works correctly.

Here’s the scenario:
Solaris pandora_agent.conf.  The agent reports every 60 seconds.
module_begin
module_name Disk Util /
module_type generic_data
module_exec df -k / | tail -1 | tr -d “%” | awk ‘{ print $5 }’
module_max 100
module_min 0
module_description root disk utilization
module_end

Linux pandora_agent.conf.  The agent reports every 60 seconds.
module_begin
module_name disk_root_free
module_type generic_data
module_exec df -kh / | tail -1 | awk ‘{ print 100-$4 }’
module_max 100
module_min 0
module_description Free disk Percentage of root partition
module_end

The Interval for both modules is 60 seconds.
Create an alert template with the following values:
Template Step 1
Priority:Warning
Condition Type: Warning status
Template Step 2
Time Threshold: 5 minutes
Default Action: send an email
Template Step 3
Alert Recovery enabled

Create an alert for the Solaris and the Linux modules, and use the same alert template for both. 
Set the Warning Status min and max in each module to a range that will cause the module to send Warning status.

Result:
If the threshold is 5-15 minutes, Pandora will send an alert email for the Solaris agent every minute, but not for the Linux agent.  The linux agent will use the Time Threshold value correctly.  If you change the Time Threshold value in the template to 30 minutes or greater, the emails for the Solaris module will stop.  If you change it back to 5-15 minutes, they will start again.

I have included XML data from a Solaris 10 agent and Linux (CentOS 5.4) below.  I’ve tested this problem with the Windows agent, and it works fine.

Solaris XML

 
   
   
   
 

Linux XML