Forum Discussion

Chris_Phillips's avatar
Chris_Phillips
Icon for Nimbostratus rankNimbostratus
Jul 13, 2007

monitoring the F5 SNMP tables

Hi,

 

 

Not too confident that this is really in the spirit of this new forum, but it's certainly about monitoring and management with the F5...

 

 

Is anyone aware of any existing products, or simple approaches to take things like the LTM virtual services SNMP table as a whole and monitor it? we're struggling to generally piece together a logical monitoring architecture where I work, and as the LTM's are monitoring the bulk of our critical services anyway and have a number of nice and bespoke SNMP tables just sitting there listing availability of nodes, services and their respective throughputs and such, should i be able to hook into this lovely data from another dashboard or something? I appreciate this is sounding vague, but i'd love to be able to directly leverage the existing monitoring that's going on rather than have to replicate it all just so other departments can see traffic lights...

 

 

(Is it just me or do *ALL* monitoring systems suck big time when you get down to the fiddly stuff?)

 

 

Thanks

 

 

Chris

10 Replies

  • Are you interested in performance data, state data, or both?

     

     

    Perf data is relatively easy to collect with almost any SNMP polling engine, the real challenge appears to be deciding what data in useful, what data aggregation gives the granularity needed w/o huge storage costs, and how to mine and display the data.

     

     

    For state (up/down type stuff), there are several companies that offer MOM management packs that can collect some of the SNMP data from the Enterprise mibs, but it's not what I'd call polished and robust. If you can hold out until the end of the year, I expect to see some much improved BIGIP monitoring solutions from the big players in Network monitoring.

     

     

    J
  • well essentially i'd like to monitor the entire table. just like a normal snmp enabled nic will have status, speed, throughput, so like i could monitor a single end server on the basic enterprise MIB, AFAIR just about all the information (well, not speed obviously...) is already available in those lovely tables, both for virtual servers and pools.

     

     

    The other angle is to monitor each virtual server automatically... i want to say i want to monitor this table. not this oid which relates to this virtual server etc... it's one nice big lump already, read it as such. Anything I've ever seen to monitor arbitrary snmp data is just one oid etc...

     

     

  • I have utilized two open source projects to achieve most of what you are looking for. I use cacti to graph the performance & trend data, and I have used Nagios (at previous employment) to send emails/pages on important state changes. In cacti, the virtual server snmp table is indexed automatically, so I don't have to do the manual oid plugging to get the desired end result. I am trying to accomplish this with the pool table, but my oids get expressed with a double dot in the middle, so I haven't quite figured out the regex for that table. If you decide to pursue the Cacti path, I can shoot you some screenshots as well as the F5 template I customized for the cacti forums. I could probably put a page together for the wiki as well and post the templates & screenshots there if anyone thinks that would be helpful.
  • I too am looking into using cacti and nagios for monitoring state changes and graphing data. If you don't mind, I would like to see what you have come up with so far.

     

     

    ty
  • I would also love to see the screen shots and template you came up with for Cacti.
  • if it's of interest to anyone, it looks like we're about to use ManageEngine Application Manager to monitor the whole table. it capable of automatically sucking up any given table and (afaik) monitor all values of any column.
  •  

    After your post I had a look at product. The knowledge shows a few topics complaining they have no or little luck monitoring bigip devices due to MIB issues. Go to main page use search "bigip".

     

     

    Have you tested it? How did it go? I'm looking for a network management (fault/monitoring/reporting) product for my company at the moment.
  • Did you get anywhere with ManageEngine Application Manager? I am using Opmanager and can not any stats with SNMP even when specifying a OID i know works in MRTG.

     

     

    Looking at icontrol/powershell as an alternative collecion method.
  • AppManager can certainly obtain the tables as part of a monitor, but then i didn't know enough about appManager when i was playing with it to go much further.

     

     

    I think part of my problems about this sort of thing is that rather than mindlessly monitoring an entire table, a more formal docuemnted process for support within an IT department needs to define what should be monitored. I shouldn't have been seeing manually associating each VS with (in this case) an application monitor on AppManager as a big thing, but the responsible way to combine the different parts of what comprises an application.

     

     

    And in fact now, i don't really know how monitoring "an entire table" is supposed to even work...