Monitoring Exchange 2003 from Linux Cacti

Templates, scripts for templates, scripts and requests for templates.

Moderators: Moderators, Developers

Post Reply
Author
Message
zeroy
Posts: 25
Joined: Wed Jan 25, 2006 8:27 am

#16 Post by zeroy » Thu Jan 26, 2006 7:46 am

Some pics of my graphs...
Attachments
1.jpg
Connections
1.jpg (52.58 KiB) Viewed 15845 times
2.jpg
Delivery Filters
2.jpg (58.41 KiB) Viewed 15845 times
3.jpg
IMF filter
3.jpg (94.69 KiB) Viewed 15845 times
4.jpg
RPC requests
4.jpg (51.16 KiB) Viewed 15845 times
¬¬ZeRoY¬¬

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#17 Post by argon0 » Thu Jan 26, 2006 8:17 am

Yup - getting the Disk space and Interface graphs

Ah... That'd be why then...

Is this easy to do? Does it require reboot of the Win2k box?

Argon0
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#18 Post by argon0 » Thu Jan 26, 2006 8:20 am

Lovely Graphs BTW - not got IMF installed, but other graphs would be very useful.

Argon0
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

cleaner
Posts: 36
Joined: Mon Feb 28, 2005 10:12 am

#19 Post by cleaner » Thu Feb 02, 2006 5:25 pm

Getting some weird stuff from the scripts. The users script works fine but all the others out put this:


02/02/2006 03:00:29 PM - CACTID: Poller[0] Host[27] DS[853] SCRIPT: perl /usr/local/apache2/htdocs/cacti/scripts/exchange_imf_script.pl sssexch01, output: U
02/02/2006 03:00:29 PM - CACTID: Poller[0] Host[27] DS[853] WARNING: Result from SCRIPT not valid. Partial Result: TotalMessagesAssign...


2/02/2006 03:00:29 PM - CACTID: Poller[0] Host[27] DEBUG: The POPEN returned the following File Descriptor 10
02/02/2006 03:00:29 PM - CACTID: Poller[0] Host[27] DS[852] SCRIPT: perl /usr/local/apache2/htdocs/cacti/scripts/exchange_filters_script.pl sssexch01, output: U
02/02/2006 03:00:29 PM - CACTID: Poller[0] Host[27] DS[852] WARNING: Result from SCRIPT not valid. Partial Result: BlockListDNSQueries...


02/02/2006 03:00:18 PM - CACTID: Poller[0] Host[27] DEBUG: The POPEN returned the following File Descriptor 10
02/02/2006 03:00:18 PM - CACTID: Poller[0] Host[27] DS[851] SCRIPT: perl /usr/local/apache2/htdocs/cacti/scripts/exchange_emails_script.pl sssexch01 get SendQueueSize,ReceiveQueueSize,MessagesSentPermin,MessagesSubmittedPermin,MessagesDeliveredPermin, output: U
02/02/2006 03:00:18 PM - CACTID: Poller[0] Host[27] DS[851] WARNING: Result from SCRIPT not valid. Partial Result: MessagesSubmittedPe...

User avatar
vedd99
Posts: 32
Joined: Wed Jan 11, 2006 4:51 pm
Contact:

connections graph

#20 Post by vedd99 » Fri Feb 03, 2006 12:18 am

any idea why the connections graph would only display active connections? when I run the script manually from the command line everything seems to be working fine, it just doesn't make it to the graph.

gentoo scripts # perl exchange_users_script.pl 10.x.x.x
UserCount:168 ActiveUserCount:107 ActiveConn:338

User avatar
vedd99
Posts: 32
Joined: Wed Jan 11, 2006 4:51 pm
Contact:

Re: connections graph

#21 Post by vedd99 » Fri Feb 03, 2006 12:39 am

vedd99 wrote:any idea why the connections graph would only display active connections? when I run the script manually from the command line everything seems to be working fine, it just doesn't make it to the graph.

gentoo scripts # perl exchange_users_script.pl 10.x.x.x
UserCount:168 ActiveUserCount:107 ActiveConn:338
actually my graphs show lower values that what the perl script reports as well, very strange (see screenshot).
Attachments
exchange-connections.JPG
exchange-connections.JPG (28.79 KiB) Viewed 15745 times

Rene Doove
Posts: 4
Joined: Tue May 04, 2004 7:57 am

#22 Post by Rene Doove » Fri Feb 03, 2006 3:59 am

I have the same problems as vedd99: script runs fine
but graph is not always showing total number of users.

anyone got an idea?
Attachments
graph_image.php.png
graph_image.php.png (5.58 KiB) Viewed 15741 times

zeroy
Posts: 25
Joined: Wed Jan 25, 2006 8:27 am

#23 Post by zeroy » Fri Feb 03, 2006 8:34 am

Its actually the same for me, not always...Could be a graph Template issue, there were designed for Cacti running on Windows box, not sure how to fix it :(
¬¬ZeRoY¬¬

User avatar
vedd99
Posts: 32
Joined: Wed Jan 11, 2006 4:51 pm
Contact:

#24 Post by vedd99 » Fri Feb 03, 2006 8:59 am

Rene Doove wrote:I have the same problems as vedd99: script runs fine
but graph is not always showing total number of users.

anyone got an idea?
I have read on another thread that you can change the Data Template value from 100 to 0 to fix your problem.

Go to Console / Data Templates / Exchange - I.S. Stats and select the count_users tab and change the Maximum Value to 0.

I tried this for mine with no luck, but I think I have a different problem with the actual graph template.

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#25 Post by argon0 » Wed Feb 08, 2006 7:14 am

Hokay, Now I've got check_nt installed, nsclient on the Exchange server (and some others) and can gather some of the data I want (e.g. System Uptime, active connections, etc..)

But not what I really want (message stats).

I understand how check_nt works, AND how the results are turned into graphs (at last) and have created a couple of graphs myself :D ...

I can run the query from a command line correctly, and get expected results, but I'm STILL getting nans on my Exchange Delivery Stats graph...

HAYELP!!!

(actually just had a thought, wonder if the data template is passing the correct variable through, I'll have a tinker and find out if I can get this to work...)
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#26 Post by argon0 » Wed Feb 08, 2006 8:16 am

AHA!!! I think it may have ben a permissions thing... check_nt in .../nagios/plugins was owned by root, check_nt in /var/www/html/scripts was owned by apache (which is what I think Cacti runs as in CactiEZ...

Changed, lets see what happens now!
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#27 Post by argon0 » Wed Feb 08, 2006 8:24 am

On the other hand I think I spotted a flaw with this method of data gathering... The counter returned is e.g. deliveries per min, BUT this is a one off query (i.e. its the CURRENT deliveries per min).

Cacti graphs every 5 minutes, so what you are getting is not the average deliveries per min over the 5 minute period BUT the Current Deliveries per minute AT THE POINT IN TIME THE MEASURE IS TAKEN.

What is needed is a similar method to that used when gathering snmp data for bandwidth.

I.E. count no. of messages delivered TOTAL, (call this instance A), 5 minutes later measure again (instance B) then subtract A from B, and divide by 5 to get no. of messages delivered per minute in the last 5.

Some things do lend themselves better to spot checks (e.g. CPU utilisation - but even this would be better done as an average over last 5 minutes - although not sure how you'd get that data)...

Any thoughts? OR have you already taken this into consideration, and I've just missed it?
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#28 Post by argon0 » Wed Feb 08, 2006 10:18 am

Nope...not JUST a permissions thing....

Looking at the Data Input method, why does it read like this:

Code: Select all

perl <path_cacti>/scripts/exchange_emails_script.pl <hostname> get SendQueueSize,ReceiveQueueSize,MessagesSentPermin,MessagesSubmittedPermin,MessagesDeliveredPermin
i.e. what purpose does the get perform?
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

argon0
Cacti User
Posts: 443
Joined: Fri Mar 12, 2004 1:22 pm

#29 Post by argon0 » Thu Feb 09, 2006 7:22 am

Ok, well I took this BACK to basics and using the perfmon_data data input method posted over with a LOAD of templates somewhere on here (search for NSClient or follow this link: http://forums.cacti.net/viewtopic.php?t ... highlight=)

Using the same variables plugged into the permon script that the exchange_emails_scipt.pl uses & your graph template - using the new Data sources I get a graph that works!!

Here's the template:
Attachments
cacti_graph_template_exchange_delivery_statistics_v2.xml
(24.66 KiB) Downloaded 1065 times
[size=50]No longer a n00by but[/size] not yet an 0ldby

Now Head of Technology at RSCH, back to the prickly subject of Monitorring....

scen
Posts: 5
Joined: Thu Feb 02, 2006 4:03 am

#30 Post by scen » Fri Feb 10, 2006 6:54 am

what is the meaning of Instance Name and Host Name for the Exchange Graphs?

I received only 0

Is the Host Name the same as the Exchange Name?

Post Reply