- observium - lists.observium.org

question regarding traffic accounting
by Ross [Eve IT] 09 Mar '17

09 Mar '17

Hello all, I'm using traffic accounting for multiple interfaces and have noticed a strange occurrence. The graphs seem to be be producing the wrong time in the graphs. See attached; [image: Inline image 1] If you notice the top right corner, 2017/03/08 03:15:04 It's Tues 7th here 5:15pm. The server clock is synced NTP. Is there something I've missed here ? Observium 17.3.8420 (rolling) Debian Wheezy. Ross.

7 25

Logs for LDAP
by David Haché 09 Mar '17

09 Mar '17

Hi, I have been using observium for a while and I am pleased with it. Now when I try to integrate LDAP, I can't find the logs to troubleshoot. I activated LDAP and it connect correctly, the query looks good but no go. Log from my LDAP server [06/Mar/2017:16:41:55.102124385 -0500] conn=255138 op=1 SRCH base="cn=users,cn=accounts,dc=domain,dc=com" scope=2 filter="(&(objectClass=posixaccount)(uid=dave))" attrs=ALL Does anyone have suggestion how to activate the loggin for ldap on the PHP / NGINX / Observium side? Thanks. Dave

3 5

Windows Processor Numbers Seem Odd.
by Don Maker 09 Mar '17

09 Mar '17

"643m means 0.643 Adam." Thank you for the response, Adam. That makes sense, but numbers like 291.59m do not. Should I just assume that anything with an m is less than zero?

1 0

Help with monitor Tomcat Traffic without Apache
by Michael 09 Mar '17

09 Mar '17

2 1

Windows Processor Numbers are Odd
by Don Maker 09 Mar '17

09 Mar '17

I got observium working almost exactly how I want it, but I have an issue with Windows cpu numbers. The graph looks fine (I think), but the numbers are weird. The max numbers look like a percentage, but the now and Avg show odd things like 640.5m I have no idea what that means. Also, the processor is never identified properly. I would happily change to wmi for Windows devices, but I'm not sure how that is done, or if it is stable. Does anyone have any suggestions or experiences with wmi and observium? Thanks!

4 3

Discovery creating & deleting inventory entries
by Lorenzo Zafra 07 Mar '17

07 Mar '17

Hi All, When I run discovery.php on our Nokia 7210 Service Access Switch (Alcatel-Lucent TiMOS) it creates and deletes inventory entries and the valid entries are found at the bottom of the page. See screenshot below. Are these deleted automatically after some time? I found that there is an inventory.inc.php file for the housekeeping script, but it is not included in the options when running housekeeping.php. Does this mean that inventory.inc.php is deprecated? To delete them, I just edited the script myself and ran it. Thanks! -- [image: Edmonton_sig_RGB_S.jpg] Lorenzo Zafra COMMUNICATIONS ENGINEERING STUDENT ENGINEERING & MAINTENANCE CITY OPERATIONS | EDMONTON TRANSIT 7 <(780)%20442-5021>80-944-0419 OFFICE 7 <(780)%20668-1302>80-885-3696 MOBILE l <matthew.dyck(a)edmonton.ca>orenzo.zafra(a)edmonton.ca City of Edmonton D.L. MacDonald Transit Yards ROW Building 13304 - 50A Street Edmonton AB T5A 4P6 All information contained in this email post is proprietary to the City of Edmonton, confidential and intended only for the addressed recipient. If you have received this post in error, please disregard the contents, inform the sender of the misdirection, and remove it from your system. The copying, dissemination or distribution of this email, if misdirected, is strictly prohibited.

2 2

Event LOG ignore
by Alexandre J. Correa (Onda) 07 Mar '17

07 Mar '17

Hello, Thre´s a way to ignore/filter some event log like this (flooding, in this case): 8h 30m 57s GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 14h 30m 37s GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 4d 14h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 4d 20h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 5d 20h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 7d 8h 31m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 7d 20h 31m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 8d 2h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 8d 8h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 8d 14h 31m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 8d 20h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 9d 8h 30m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:9f:26:47 -> 00:0c:29:51:1e:76 10d 8h 31m GE0/0/5 MAC changed: xxx.xxx.xxx.xxx : 00:0c:29:51:1e:76 -> 00:0c:29:9f:26:47 .... .... some servers share same loopback address (dns anycast). regars, -- Fwd: a *Alexandre Jeronimo Correa* / CEO _alexandre(a)onda.net.br <mailto:alexandre@onda.net.br>_ / Office +55 34 3351 - 3077 *ONDA INTERNET* +55 34 3351-3077 Av. Benedito Valadares, 217 – Centro – Sacramento – MG - BR _http://www.onda.net.br <http://www.onda.net.br/>_

3 2

artificially high (100%) CPU usage reported from routing engine on Juniper MX-104 w/ 550 interfaces & solution
by Darrell Budic 07 Mar '17

07 Mar '17

Observium has been recording artificially high CPU for RE0 on my MX-104, and I finally spent the time to understand what was going on and look for a solution. Here’s what I found in case it helps someone else. I say artificially because monitoring the system from the CLI reported 30-40% CPU usage with significant spikes during SNMP polling, but Observium constantly recorded 100% CPU usage. First up, I got the juniper to stop reporting on it’s internal interfaces and a bunch of others that show up by default. This removed a bunch of interfaces that existed but never had data and observing was polling. This cut the poll time down a fair bit, so I mention it if you haven’t used it on a juniper before. If you want this, it’s: snmp { filter-interfaces { all-internal-interfaces; interfaces { <other interface you don’t want stats for here>; } } } I also went through and turned off all the modules it wasn’t using or weren’t producing any data in polling or discovery. Helped a little more, but exposed something odd. When I would run polling on just the processors module, it would report CPU usage that matched what I saw in the CLI. When it ran as part of the whole host polling, it reported 100%. Figured out it was polling the processors after it had polled sensors and the 550 interfaces on this router, and that has run up the CPU utilization while it was polling, so the RE was reporting that instant CPU usage while it was still high. Not sure it was the best way, but moved the processor module to the head of the polling queue, and now I’m getting a CPU status that is more representative of the general load on the RE, and not the 100% usage the SNMP polling was causing. This router has always had this many interfaces, so as far as I can tell, a change in polling order caused this behavior and not a growth in interfaces. Here’s my patch to includes/polling/functions.inc.php if it’s helpful to anyone: Index: functions.inc.php =================================================================== --- functions.inc.php (revision 8345) +++ functions.inc.php (working copy) @@ -639,6 +639,11 @@ //print_warning("Module [ $module ] excluded for device."); continue; } + if ($module == 'processors') + { + array_unshift($poll_modules, $module); // Add 'processors' before all + continue; + } if ($module == 'unix-agent') { array_unshift($poll_modules, $module); // Add 'unix-agent' before all

1 0

migrating observium to a new server
by Shawn Faulkingham 06 Mar '17

06 Mar '17

Is anyone aware of any documentation to migrate an observing installation to a new server? I am moving from CentOS 6 to CentOS 7 for security and PHP reasons. If anyone has any pointers, I would greatly appreciate it. Thanks, Shawn Faulkingham

4 8

Observium-OpsGenie integration
by Celal Emre Çiçek 06 Mar '17

06 Mar '17

Hello, My name is Celal Emre ÇİÇEK and I'm an integration engineer at OpsGenie. I'd like to know if there are any changes in the latest updates of Observium that could affect Observium - OpsGenie integration. Today, one of our customers came up with a problem. He says Observium alerts aren't being forwarded to OpsGenie. He also says that he didn't change any configuration in Observium and in OpsGenie. We've confirmed that he didn't change any configuration in our side. We also asked him to check his Observium configuration and compare it with our docs. He's checked and nothing is wrong. So, we wonder if there are any changes that affect OpsGenie integration in Observium in the latest updates, maybe some changes in the backend. Could you please check and return to us? Best regards, Celal Emre ÇİÇEK

2 2