Kaseya Community

Disk/Memory/CPU Alerts

Posted by LegacyPoster

on Mar 20, 2010 3:34 AM

I have spent a lot of time fine-tuning many of our alerts and it has resulted in fairly efficient alerting in a number of areas, including disk space, services, anti-virus, etc.

Where we have struggled is some of the performance alerts. We have found some of the counters don't work properly or we have applied them improperly because it wasn't clear what they were designed for. Others generate excessive alerts and seem to be fine by the time we check them. Here are some areas we wish to get alerts on:

Excessive disk transfers/sec.
Excessive CPU usage
Excessive page file usage
Excessive RAM usage

These are not so important on desktops, but could be very useful on servers. Any ideas on how to best monitor these? Obviously, we'd have to fine tune the values to fit various scenarios, but we'd like to know the best ways to monitor these performance issues.

Legacy Forum Name: Disk/Memory/CPU Alerts,
Legacy Posted By Username: bjfreeman

Posted by LegacyPoster

on Mar 29, 2010 9:37 PM

We monitor the "Logical Disk Avg. Disk sec/Transfer" to determine how busy the disk system is. This value tells you how long the system needs to wait before it can transfer data to/from the disk. It is the best means I have found to measure "busy" disks regardless of their RAID status. Many of the other values gets thrown off or useless once you have a RAID system.

The alarm threshold is set to 0.03 for 30 minutes with sampling every 10 minutes. So, three consequtive samples > 0.03 will trigger an alarm.

For CPU, the alarm threshold is over 85% for 30 minutes.

We don't really monitor the others, as these would be very individual for each server monitored. I guess if available physical memory dropped below a certain amount, that could be considered "bad" and warrant a look...

Legacy Forum Name: Monitor Sets,
Legacy Posted By Username: Lmhansen

Posted by LegacyPoster

on Mar 29, 2010 11:07 PM

If you need to monitor page file usage, the Memory\% Committed Bytes in use appears to be the way to go. It give you Memory\Commit Limit / Memory\Committed Bytes, in percent. Once that number creeps up to 100%, then the pagefile will grow (assuming it can).

Legacy Forum Name: Monitor Sets,
Legacy Posted By Username: Lmhansen

Posted by LegacyPoster

on Jul 7, 2010 12:33 AM

Hello,
I starting working with Kaseya 5.1 last week, maybe this is very basic so I apologize for that. But I have been through the KB and I haven't found information regarding the counter values.

I try to monitor CPU utilization with Monitor> Monitor Sets
pull the report from Reports>Monitor

But I have problems interpreting the counters values, I should be able to explain the CPU usage to the client using this counters graph.

I am setting the following values in my lab environment :
Object: Processor
Counter : %Processor Time
Instance: _Total

Collection Threshold :
* Collection Operator : Over
* Collection Threshold: 1%
* Sample Interval: 5 minutes

Alarm Threshold
* Alarm Operator: Over
*Alarm Threshold: 5
*Duration: 10 Minutes

Ignore aditional alarms for : 1 hour

Warn when within X% of alarm threshold: 20
Trend activated: No

with what values are you working with?

I would appreciate any information that help me to understand this.
Thank you.
Ana

Legacy Forum Name: Monitor Sets,
Legacy Posted By Username: anbant

SQL Queries for Cpu , Disk, and Memory Utilization

CPU, Disk, Network, Memory in Live Connect Sessions...

Database table name for CPU,disk, and Memory Utilization?

Are you monitoring CPU/Memory/Disk Usages to get live graphs

CPU/ Memory alarms

Disk/Memory/CPU Alerts