Looking for suggestions on what to monitor with a 3rd party monitoring

[AaronAllen]

My company is currently looking at writing up a project to purchase an enterprise wide monitoring tool. I'm personally responsible for the UNIX side of the operating systems. So far I haven’t found any site that can give me a list of everything that can be monitored within AIX. So just to get the process started I've come up with the basic things as follows.

1. Disk
a. % full. (Able to run a script if % = specified number).
b. Growth. (Monitor, track and trend spikes and normal growth).
c. Able to monitor at the file system level.

2. CPU
a. Monitor total % utilization.
b. Monitor CPU by individual % utilization
c. Utilization (Monitor, track and trend spikes and normal growth).

3. Memory
a. Monitor total % utilization.
b. Growth. (Monitor, track and trend spikes and normal growth).
c. Monitor paging.
d. Monitor paging faults.

4. Processes
a. Able to monitor custom processes as defined by application support for % CPU utilization, alive/dead and memory utilization.
b. Able to run scripts based on results of monitoring.

5. Logs
a. Able to parse logs for predefined search criteria and notify or (track and trend) results.

6. Network
a. Able to monitor network utilization by protocol or port.

This is all I have so far, but I'd like to get feedback on anything anyone thinks I should add.

Thanks,

 

[Mag0007]

Tivoli monitoring by IBM/Tivoli is nice...

 

[Germo]

Take a look at Nagios Aaron. It runs on a Linux server and can monitor any server Unix or NT/Windows.

http://www.nagios.org/

Hope that helps

 

[hirschaj]

Just to make note, 3a and 3b will be pretty useless in AIX since memory usage grows toward 100% on any server that is running for a long period of time. 3c is the really important one. Not sure what you are looking for with 3d, that one should vary greatly by application.

I would also want to monitor for network errors (crc, dma, collisions, no receive pool buffers).

You didn't mention print queue monitoring but it might apply.

You might also want to monitor FS buffer errors.

Whatever product you choose should be able to have custom modules plugged in for your particular needs.


Jim Hirschauer

http://www.aixexpert.com

 

[AaronAllen]

Hirschaj,
Thanks for the information that type of information is what I am looking for. We aren't choicing a product as of yet, just getting the requirements down for the verdors.

 

[hirschaj]

I forgot something really important in my last post.

Having a tool that can parse the error report for hardware errors can be very helpful. You really want to get an alert when one of your mirrored disks dies and you are left with 1 good copy. If you are buying a product it might as well have this funtionality instead of writing your own scripts.


Jim Hirschauer

http://www.aixexpert.com

 

[hfaix]

I use "BigBrother" or you could try "BigSister". It's not that best and you can write about anything you want for it, but you can't beat the price (free or reasonable depending on your usage).

I typically monitor,

ftp, ping, snmp, telnet,
processes like cron, auditbin, and others app processes
database logs,
errpt
cpu (sar)
paging
disk

disk, paging, cpu are graphed as well.


and other stuff that I can't come up with off the top of my head.

it can be configured to page pagers (which I don't use), email me, email the datacenter... or email who you want.

 

[ereshatezer]

Did you try nmon? Most of the mentioned things are covered.

http://www-128.ibm.com/developerworks/eserver/articles/analyze_aix/

 

[plamb]

use HOBBITMON instead of BigBrother looksk thesame but much faster