Powershell monitoring Connected Mobile Experience - arrays

CalasTyph0n

Hi,

so I'm trying to hack together some rudimentary monitoring, I have the below script:

$wd = 'c:\CMX\'
$files = gci $wd
$pwd = pwd
$alerts = "alerts.csv"
$tmp = 'tmp.txt'

cd $wd

foreach ($l in $files){
$ip = $l.BaseName | %{ $_.split('-')[1]; }
$count = get-content $l | %{ $_.split(',')[0]; } | %{ $_.split(':')[1]; }
(Write-Output $ip.Trim() $count.Trim())-join "," >> $alerts
}

# Alerts
$alert = Get-Content $alerts
$chost = @()
$calarm = @()

foreach ($a in $alert){
    if($a -notmatch ".*,[0].*"){
        $chost+= $a.Split(',')[0];
        }
    if($a -notmatch ".*,[0].*"){
        $calarm+= $a.Split(',')[1]; 
        }
}

write-host "Message.CMXAlertCount:" $chost[0]
write-host "Statistic.CMXAlertCount:" $calarm[0]


rm $alerts
exit 0;

which when run in powershell returns:

array index 0:

Message.CMXAlertCount: 172.16.4.10
Statistic.CMXAlertCount: 1

array index 1:

Message.CMXAlertCount: 172.26.4.10
Statistic.CMXAlertCount: 2

but the above is manual, is there a way to loop it? so dynamically say for each index in the array perform the write host?

if there is, how do i have solarwinds take that output as i believe it will error.

I am querying multiple systems and dumping that data into one file that contains the host ip and the number of active alerts on that host, all i want to be able to do is monitor this file and when an alert value increases trigger something, be it an email or syslog whatever.

this is my first time using SAM so am probably going about it the wrong way, i should add for context the systems are remote and currently solarwinds does not have direct access and is unlikely too.

thanks in advance.

mesverrum

To your final point that you are using SAM "wrong" that is kind of true.  SAM is written with the assumption that it will be the tool scraping all your servers for data and then reporting on them separately.  The tool is built around that so doing what you want to do is going to require some kind of work around.  If you hadn't added the note about the restricted access I would tell you to just hit the hosts directly to make life easier, but if we can't then we have to get creative.

A single script component does allow you to return up to 10 different statistic/message pairs per component, you just have to give them different names in the format of something like Statistic.$chost: $calarm (and in this case you could probably ditch the messages completely since the statistic will actually end up with the label of the host already).  Those names basically need to be persistent in terms of their label and the number of them, which becomes a headache for us to work around.  So you can't have it dynamically generating them on the fly where it changes each time.  You CAN dynamically generate them the first time the script gets tested but SAM then basically sets up some placeholders in the DB and will just be looking for the one possible outputs that it already knows about.  This part probably doesn't make much sense right now but once you start using scripts with multiple outputs in SAM it will be a problem you frequently run into and you learn to watch out for it.

Since the script monitors don't like it if you have a dynamically changing set of outputs, for your scenario assuming there isn't a crazy number of hosts you might just make enough copies of the component so that you could split up all your hosts into sets of 10, then pass the list into each component as an argument so it would be like

Scripts Arguments: ip1,ip2,...,ip10

Then you'd just have something like a where-object filtering your $ip variable to only take in the ones that are part of the array @args.

To do the looping it looks to me like you'd just move the write-host chunk inside of the $alerts foreach loop and modify your write-output like I mentioned before, Statistic.$chost: $calarm.

Then you have to test the component with each set of args and you will see SAM create those placeholders I was mentioning and then you should be good.

You can see one example of a multi-output script I made here. In this example you'd see that I defined the total and process 1-5 as my pre-tested outputs.  I don't think that you would be so lucky as just being able to do the 1-5 shorthand because in your case it looks like you probably want to track the hosts and their errors over time, in my case I just needed an option to spot check the names of the procs in the top 5 slots and knew they would be continuously changing from host to host and poll to poll.

https://thwack.solarwinds.com/t5/SAM-Documents/CPU-MEM-Process-Snapshot/ta-p/528811