Varnish 3.0.3, CentOS 6.3 - how to properly monitor it with Nagios?

Rafał Radecki radecki.rafal at
Tue Mar 19 13:51:50 CET 2013

Hi All.

I am searching information on how to properly monitor varnish
parameters with Nagios+nrpe.
I would like to monitor backend health and various parameters
(client_conn, n_lru_moved, ... ). I know that I can use:

- -> but it only gives me
cumulative number for every parameter

./check_varnish -p client_conn
VARNISH UNKNOWN: Client connections accepted (8,361)|client_conn=8361
and not a per second value which is what I would like to have;

- varnishstat -1 -> but it also gives me only cumulative number

varnishstat -f client_conn -1
client_conn               8409         0.49 Client connections accepted
(0.49 - the average since varnish startup also not usable, rather
infromational only).

I also know that there is no stats command in varnishadm for varnish 3
so I cannot use it.
I would like to monitor my parameters' per second (or other timeperiod
defined) change, do you know any solutions/scripts/tools?
Is my only option to write a script which (for example) calls
varnishstat -1 every second for some time and then computes the

Best regards,
Rafal Radecki.

More information about the varnish-misc mailing list