01-16-2013 05:01 AM
what is the use of "Uptime" counter in Unix resource Monitor in sitescope 11.20 ?
Thanks and regards,
Solved! Go to Solution.
01-16-2013 05:44 AM - edited 01-16-2013 05:50 AM
Looking at the config file ($SITESCOPE_HOME\templates.os\browsable\Liux26Comm
I suppose what you'll see in Sitescope is how long the target machine has been running since it was last started up/ rebooted. Unfortunately I can't confirm this as the machine and OS account I used to test with doesn't have execute permissions for /proc/uptime.
Pretty sure it will just show you how long the target machine has been running for uninterrupted.
EDIT: Just to give you an idea of what the counter does in the background, here's an extract from the config file:
<command name="Uptime" class="FileParser">
<arg type="map" name="command">
<element name="exe" value="cat"/>
<element name="options" value="/proc/uptime"/>
<arg type="map" name="match1">
<element name="MatchValue" value="\d+.*\s+\d+.*"/>
<element name="InstanceName" value="null"/>
<element name="ObjectName" value="Uptime"/>
<element name="Counter:Uptime" value="group:1"/>
Here's what you typically get when you run the "uptime" command straight on the OS:
15:44:54 up 1 day, 22:55, 11 users, load average: 83.42, 86.03, 85.00
Looks like it might show the whole line in Sitescope, but it might filter it more to show only the uptime value in days.
01-16-2013 10:58 PM
I can't verify this on my side unfortunately, but I suspect it will likely give you the uptime in days. Easy way to confirm is to configure the monitor for a particular target server and see what counters are returned in Sitescope (also make note of the values each counter returns), then SSH to the actual target server and run the "uptime" command manually.
By comparing what's returned in Sitescope and what's returned by the command itself, you should be able to see exactly what Sitescope measures.
I'd personally probably have followed up with the target machine's relevant sysadmin/ owner and find out if they want to be alerted if the uptime in days/ hours dip below a certain value, and if so, what value exactly. Starting with alerting if uptime in days is, say, == 0 is probably a good place to start. As soon as the uptime value is exactly 0, that essentially means that the server in question was rebooted/ shut down.
I think this monitor may cause some redundent alerting if the machine in question gets rebooted from time to time on purpose, but in that case the alerts can always maybe just be temporarily disabled beforehand.
I think it may also be used purely for informational purposes (i.e. no thresholds or alerting set). Someone might just like to see what the readings were in a weekly report or something.
Lots of options around this. If you need any further help, let us know and we'll go from there.