[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Only one alert

From: David Jones
Subject: Re: Only one alert
Date: Sun, 1 Sep 2019 13:06:25 +0000

Since you are launching a script from monit, how about this approach?

check program webserver_vhosts with path "/usr/local/bin/" every 4 cycles
    if status > 0 then alert

Then in the there could be something like this:


VHOSTS="url1 url2 url3 url4"

for VHOST in $VHOSTS; do
  /usr/lib64/nagios/plugins/check_http -H $VHOST -t 20 -f follow -u "/" -s Copyright
  [[ "$?" -ne 0 ]] && ((ERRORS++))

exit $ERRORS

The key to what you are needing to do requires a web client that will send the host header to your web server.  This could be wget, curl or a number of tools but it's more than a simple TCP port check.

That command above looks for the word "Copyright" in the web server output so adjust as needed.

Obviously this is using the nrpe-plugins tool check_http so install that package.

From: monit-general <monit-general-bounces+djones=address@hidden> on behalf of Paul van der Vlis <address@hidden>
Sent: Saturday, August 31, 2019 3:55 PM
To: address@hidden <address@hidden>
Subject: Re: Only one alert
Hello Lutz, and others,

Op 30-08-19 om 22:01 schreef Lutz Mader:
> Hello,
> the answer is no, I think.
>> I want to monitor several sites on the same host, both http and https,
>> but I want to get only one alert when there is a problem. Is that possible?
> But you can use a dummy service with multiple failed host tests.
> check process Server matching "sleep"
>   start program "/bin/ksh -c 'sleep 31536000 &'" with timeout 120 seconds
> #  stop program "" with timeout 120 seconds
>   if failed host port 80 then alert
>   if failed host port 80 then alert
> :
>   if failed host port 80 then alert

I must say I do not understand what you are doing here.

In your example you are using sites on different hosts. I want to
monitor more sites (virtual hosts) on the same host. But when e.g.
Apache is down, I don't want to get many alerts.

> This will reduce the number of alerts, but you get an alert if one of
> the hosts is not reachable. But I prefere the remote host check for all
> hosts and use an additional tool to handle the notification requests.

I am also executing a script.

Can you tell more about what you do to avoid getting many alerts?

With regards,

Paul van der Vlis Linux systeembeheer Groningen

reply via email to

[Prev in Thread] Current Thread [Next in Thread]