tree f8df7818de76e8062cad398edc1461f07383b12d
parent 8776f1b3a0260c7b4a5970bf0fbaf4c94f710f7d
author Mathieu Trudel <MTrudel@mediagrif.com> 1287760128 +0200
committer Willy Tarreau <w@1wt.eu> 1288458275 +0200
encoding latin1

[CONTRIB] add templates for Cacti.

Hi,

I've attached the templates I've built for monitoring backends and
frontends of haproxy.

To install these, you will need to copy the XML files from the contrib/
directory of the haproxy distribution into a directory that Cacti can
reach, and edit the Data Queries "HaProxy Backends" and "HAProxy
Frontends" accordingly (the "XML Path" field. It's also dependant on
having a version of net-snmp that supports embedded Perl, and including
the "perl do 'path_to_haproxy.pl';" directive in your snmpd.conf file.

As for what is created:

- For the devices, you have two new data queries to choose from, they
can be added from the Devices page for each device, at the very end in
the drop-down box, then click "Add". The data queries are called
"HaProxy Backends" and "HAProxy Frontends".

- From "HaProxy Backends": in the new graphs page, you can choose which
backend to graph, and create one of two graphs:
	- Haproxy backend traffic:  ingress and egress bytes.
	- Haproxy backend sessions:  total sessions with _reponse_ errors.

- From "HAProxy Frontends": in the new graphs page again, you can choose
which frontend to graph, which will include aggregated data for the
backends behind it, obviously. You can create one of two graphs:
	- Haproxy frontend traffic:  ingress and egress bytes.
	- Haproxy frontend sessions:  total sessions with _request_ errors.

In the graphs and data sources, limits are set to reasonably high values
to support up to nearly 10G traffic, and up to 10000 concurrent
connections.

/ Matt
(cherry picked from commit f63090f2e85cdb7448071cdceb2eb5fabd2b9320)