F5 BigIP v9 v10 v11 Templates (Global and LTM)

[niobe]

Huge set of templates here for BigIP v10. Broadly I've tried to pull out most of the useful stuff from the Global and LTM MIBs. Graphs included are:

• F5 BigIP - Global - Connection Rate
• F5 BigIP - Global - Connections Current
• F5 BigIP - Global - HTTP - Requests
• F5 BigIP - Global - HTTP - Responses 2xx
• F5 BigIP - Global - HTTP - Responses 3xx 4xx 5xx
• F5 BigIP - Global - IP Traffic
• F5 BigIP - Global - Packets - Client
• F5 BigIP - Global - Packets - Errors
• F5 BigIP - Global - Packets - Server
• F5 BigIP - Global - TCP Traffic
• F5 BigIP - Global - Traffic - Client
• F5 BigIP - Global - Traffic - Server
• F5 BigIP - Global - UDP Traffic
• F5 BigIP - Interface - Errors
• F5 BigIP - Interface - Packets
• F5 BigIP - Interface - Packets Multicast
• F5 BigIP - Interface - Traffic
• F5 BigIP - Member - Connection Rate
• F5 BigIP - Member - Connections Current
• F5 BigIP - Member - Packets
• F5 BigIP - Member - Traffic
• F5 BigIP - Pool - Connection Rate
• F5 BigIP - Pool - Connections Current
• F5 BigIP - Pool - Packets
• F5 BigIP - Pool - Traffic
• F5 BigIP - SSL - Connection Rate
• F5 BigIP - SSL - Current Connections
• F5 BigIP - SSL - Decrypted Traffic
• F5 BigIP - SSL - Encrypted Traffic
• F5 BigIP - TMM - CPU Cycles
• F5 BigIP - TMM - Memory
• F5 BigIP - VS - Connection Rate
• F5 BigIP - VS - Connections Current
• F5 BigIP - VS - Packets
• F5 BigIP - VS - Traffic

In all there's:

• 73 data templates
• 35 graph templates
• 4 data queries (Interface, LTM VS, LTM Pool, LTM Pool Member)
• 4 new RRAs
• 1 host template

A few things worth noting:

• Developed on 0.87g so I'm not sure if this will work at all on 0.88 - please reply with results
• Only tested on a viprion system running 10.2.3 (some or even most graphs may work on v9.x and v11.x)
• Painstaking effort made on consistent look and terminology!
• Many of the oids in the MIB returned empty for my system and these are mostly not included here. YMMV depending on your model.
• 'Current Connections' graphs - are from GAUGE oids. The give you a snapshot of the current number of active connections at the moment of polling
• 'Connection Rate' graphs - are from COUNTER oids. They give the average rate of new connections per second and so reflect to some extent what happened between polling cycles
  (the old version just had one version and said 'connections' which was a little unclear)
• The pool member graphs display IP addresses in HEX and this is UGLY. The only workaround is to change the graph names to something sensible after creation.
• The RRAs I use had to be included rather than the default cacti ones. These are optimised for a 1 min poller and a much longer and more detailed history. They result in MUCH larger rrd files that the cacti defaults. If you want to change the RRAs used by my data templates you can do it in the sql quickly and easily. Ask me if you want help with that.

Pool Member - packets per second in/out

eg.png (61.94 KiB) Viewed 33513 times

I think these are a good improvement over the previous v9 graphs but needs some testing on various platforms. So please let me know.

N

[alex500]

Thank you very much!!!

[alex500]

One question please , when I try import templates i select cacti_host_template_f5_bigip_v1020120601.xml and it appears one error message " Error:XML:Hash version does not exist" ,¿Why is it? What I do wrong ?

[niobe]

One question please , when I try import templates i select cacti_host_template_f5_bigip_v1020120601.xml and it appears one error message " Error:XML:Hash version does not exist" ,¿Why is it? What I do wrong ?

That sounds like the error you get when running a different version of cacti. Which one are you using?

[alex500]

I am using Cacti 0.8.7 e

[niobe]

Well a quick search of the forums suggests this is because of your cacti version. If you upgrade to at least 0.87g it should work.

[ejcacti]

Hi,

First of all - thanks very much for the v10-templates!

I've got them running under Version 0.8.7h by importing
"cacti_host_template_f5_bigip_v10 20120601.xml" followed by all the other ".xml"s from your zip-file via the gui.


The only issue I'm having is this:

When I add a new host and go for "Create Graphs for this Host" I'm getting "Error in data query." for

Data Query [F5 BigIP - Interface Statistics]
Data Query [F5 BigIP - LTM Member Statistics]
Data Query [F5 BigIP - LTM Pool Statistics]
Data Query [F5 BigIP - LTM VS Statistics]

I'm enclosing a screenshot sou you get the idea of how this looks like.

Thanks much in advance for any clue,
-ewald

[niobe]

It looks like some or all of the data queries parts are missing.

Did you copy the f5_* xml files to <path_to_cacti>/resources/snmp_queries?

Code: Select all

$ pwd
/var/www/cacti/resource/snmp_queries
$ ls -1 f5_bigip*xml
f5_bigip_if.xml
f5_bigip_ltm_member.xml
f5_bigip_ltm_pool.xml
f5_bigip_ltm_vs.xml

Are the data queries listed in the console?

data queries

cacti.PNG (19.73 KiB) Viewed 33380 times

Can you add the data query to a host (or a host to a host template) and do you get results from verbose query?

[ejcacti]

It looks like some or all of the data queries parts are missing.

Did you copy the f5_* xml files to <path_to_cacti>/resources/snmp_queries?

Code: Select all

$ pwd
/var/www/cacti/resource/snmp_queries
$ ls -1 f5_bigip*xml
f5_bigip_if.xml
f5_bigip_ltm_member.xml
f5_bigip_ltm_pool.xml
f5_bigip_ltm_vs.xml

Hi,

Well sure, I missed this completely.

Now I've got all the xml-files in place, don't get any of the errors mentioned above, but unfortunately all F5 related graphs only show up empty.

Whichever graph I select, be it "F5 Global - Connection Rate" or "F5 TMM - CPU Cycles" I only get "nan" values.

The cacti log file doesn't hold any meaningful information, esp no error messages.

Also note that cacti can contact the F5-box in question since I get the system status including hostname, uptime etc. displayed for the device.

The box in question runs BIG-IP 10.2.3 Build 112.0 Final; my cacti-version is 0.8.7h.

Any ideas as to why I don't get any F5-related data into Cacti?

Thanks much for your help,
-ewald

[niobe]

For that you should just follow the usual debugging NaNs procedure. Search the site, or I believe a link is always in Gandalf's signature. You can also do a manual snmpwalk - look at the data templates or xml files to get the oids.
By the way you didn't mention any data query graphs. Do verbose queries work on the data queries?

[ejcacti]

You can also do a manual snmpwalk - look at the data templates or xml files to get the oids.
By the way you didn't mention any data query graphs. Do verbose queries work on the data queries?

Hi,

Looks like some OIDs simply aren't there.

First if all I did an snmpwalk just for plain "system" -> works; so basic SNMP is ok.

Next I went for the interface statistics: took a look at "f5_bigip_if.xml" and tried some of the OIDs herein, e.g. Interface names

Code: Select all

 snmpwalk -v 1 -c ... .1.3.6.1.4.1.3375.2.1.2.4.4.3.1.1
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.49 = STRING: "1.1"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.50 = STRING: "1.2"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.51 = STRING: "1.3"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.52 = STRING: "1.4"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.53 = STRING: "1.5"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.54 = STRING: "1.6"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.55 = STRING: "1.7"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.49.46.56 = STRING: "1.8"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.50.46.49 = STRING: "2.1"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.50.46.50 = STRING: "2.2"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.50.46.51 = STRING: "2.3"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.3.50.46.52 = STRING: "2.4"
SNMPv2-SMI::enterprises.3375.2.1.2.4.4.3.1.1.4.109.103.109.116 = STRING: "mgmt"

However when going for actual Bytes by

Code: Select all

snmpwalk -v 1 -c ... .1.3.6.1.4.1.3375.2.1.2.4.4.3.1.3

I get absolutely nothing. The same holds true for e.g. Packets, Errors etc. This seems to be the reason for all those "nan" values. Could it be that the OIDs queried by Cacti don't correspond to the OIDs the F5 actually holds.

Any ideas as to what can be wrong here?
-ewald

[niobe]

Could it be that the OIDs queried by Cacti don't correspond to the OIDs the F5 actually holds.
-ewald

Right, I mention this in the first post..

Many of the oids in the MIB returned empty for my system and these are mostly not included here. YMMV depending on your model.

You are running a very similar version to me so I imagine this is a platform issue, and I would guess you're not running a Viprion. Vendors unfortunately are not always consistent with their snmp support. So a solution..

I would bet that this information is available but just in a different part of the MIB. While I was trawling through it I noticed a lot of duplication of the same information in different parts of the tree. For example interface bytes was in at least 3 places from memory. It may be that some of those alternate oids work for you. I would tackle this by firing up MIB Browser (from ireasoning), load the MIBS (download them from your bigip welcome page), and start searching for working versions of the oids I've got in my templates. MIB Browser is a great (and free) product so shouldn't take too long to find them if they're there. Then you just have to change the oids in the xml and data templates.

[ejcacti]

You are running a very similar version to me so I imagine this is a platform issue, and I would guess you're not running a Viprion. Vendors unfortunately are not always consistent with their snmp support. So a solution..

I would bet that this information is available but just in a different part of the MIB. While I was trawling through it I noticed a lot of duplication of the same information in different parts of the tree. ...

Hi niobe,

In the meantime I figured out what the problem was: I configured cacti to query for snmp v1, whereas some variables in the MIBs are 64bit counters. As soon as I changed the device-definition in cacti to use snmp v2 everything worked

So as a hint for others probably running into the same problem: Make sure you run snmpv2 when querying an F5

Thanks for your help!

-ewald

[niobe]

Ah good one, thanks for the update.

[bsdboy]

Wow, very nice job. You're my new favorite person!

They work great on v11 by the way, put them in yesterday and there are so many awesome graphs.