Many desktops

m0n · Post by **m0n** » Wed Oct 20, 2004 3:17 am

Hello,

is it possible to use AHM to check if 15000 desktops are up & running ?
If so, how ?

Thx,

m0n

KS-Soft · Post by **KS-Soft** » Wed Oct 20, 2004 10:59 am

What about simple Ping test method?

Regards
Alex

m0n · Post by **m0n** » Thu Oct 21, 2004 2:11 am

but ping test is not enough really to see if they are up & working for the users

hamoja · Post by **hamoja** » Thu Oct 21, 2004 5:24 am

You really need to be more specific as to what want monitored or define "working". Ping test is the only test that checks up/down status for PC comminucation to the network.

FLynch · Post by **FLynch** » Fri Oct 22, 2004 9:15 pm

How about basic monitors on;

Ping test
CPU
Memory
Disk space

.....for each workstation? However 15,000 * 4 monitors would be a hell of a lot for any system to poll, collate and report on.

KS-Soft · Post by **KS-Soft** » Sat Oct 23, 2004 2:26 pm

Yeah. If m0n wants to check 4 parameters of each system every 5 min, 15000*4/60/5 = 200 checks per second. HostMonitor will not start so many threads. But it will be possible to check each system every 10 min... if network is good.

Anyway, I agree with hamoja - m0n should be more specific. We don't know what exactly he wants to monitor. May be all systems located in the same network, or may be he wants to monitor hundreds of networks all over the world...

Regards
Alex

Marcus · Post by **Marcus** » Mon Oct 25, 2004 1:27 am

My checks would be:
Ping and "lanmanworkstation" service with the ping test as the master.

The biggest performance problems are dependent on your reporting. We see that when a lot of tests change status, sometimes reporting will boost cpu usage to 100%

m0n · Post by **m0n** » Mon Oct 25, 2004 7:29 am

Oki thx, ill just monitor the servers instead of all desktops

m0n

KS-Soft · Post by **KS-Soft** » Mon Oct 25, 2004 10:33 am

The biggest performance problems are dependent on your reporting. We see that when a lot of tests change status, sometimes reporting will boost cpu usage to 100%

You are generating reports using "Generate reports" action? Probably you can setup HostMonitor to create reports by timer (e.g. every 3 min)?

Regards
Alex

Marcus · Post by **Marcus** » Tue Oct 26, 2004 1:30 am

Probably you can setup HostMonitor to create reports by timer

Which we do, but also for every status change. Which will generate a lot of reports with (for example) network failures

KS-Soft · Post by **KS-Soft** » Tue Oct 26, 2004 1:42 pm

H'm... don't see good solution that can resolve this problem.
On the one hand we can implement some reports pool and create report after 1-3 sec delay - create single report for all requests that come within this time interval.
On the other hand it can effect action profiles behavior... E.g. when action profile should create report and then send report by mail

Regards
Alex