is there any way to check the website availability regularly

catiewong · 01-17-2019, 12:39 AM

my company site is running apache 2.4 on centos , sometimes have user complaint it can not access the website , but I didn't find error from log .

I think there are some website checker may check availability , but what I want is regularly check without input the URL manually .

would advise is there any good way , for example , writing a script to check http connection return 200 or not

thanks

berndbausch · 01-17-2019, 01:03 AM

Create a crontab entry with

Code:

curl -i YOUR_URL

and send yourself an email if you get a status code other than 200.

The -i option displays headers and the status code.

Alternatively:

Code:

wget YOUR_URL

An exit code other than 0 indicates a problem. The error status code (404 etc) is printed as well.

TenTenths · 01-17-2019, 04:49 AM

Sign up for https://nodeping.com/

l0f4r0 · 01-17-2019, 05:44 AM

You can use Website-Watcher (https://www.aignes.com/) as well.
Of course, it's *really* worth it if you have *many* other tasks/websites to give it otherwise it would be like killing a fly with a cannon

catiewong · 01-17-2019, 07:06 PM

Quote:

Originally Posted by l0f4r0

You can use Website-Watcher (https://www.aignes.com/) as well.
Of course, it's *really* worth it if you have *many* other tasks/websites to give it otherwise it would be like killing a fly with a cannon

yes , I know there are paid software may do that , is there any good way is free eg. the above suggestion create wget curl cron job to check ?

l0f4r0 · 01-18-2019, 04:04 AM

^ Indeed it is a paid software. I've just provided it in case you didn't know about it because it can really be so useful/indispensable for any advanced website monitoring, especially when looking for specific keywords. It's a must-have in a competitor/informational monitoring...
I don't know the basic internal commands this software relies on though...

This being said, you can absolutely meet your requirement with simple curl/wget requests with a cron as already provided.

Turbocapitalist · 01-18-2019, 04:13 AM

the curl option would be a capital -I or spelled out as --head to just check the headers. That will reduce the load on the server and cost lest bandwidth. The exit code is 0 if everything is OK.

Code:

while curl --silent --head http://www.example.com/home/ > /dev/null; 
do 
        sleep 360; 
done; 

aplay /usr/share/orage/sounds/Boiling.wav; 
DISPLAY=:0.0 xmessage -center "web site 'www.example.com' is down $(date)"

Or if you are getting into more serious numbers of servers and services then look at tools like Zabbix

TenTenths · 01-18-2019, 04:54 AM

Quote:

Originally Posted by catiewong

yes , I know there are paid software may do that , is there any good way is free eg. the above suggestion create wget curl cron job to check ?

The problem with monitoring internally is that it only goes part of the way. For example, if you are monitoring a server from itself then that doesn't take in to account anything like the external network or internet connection. Hence why I mentioned NodePing, they attempt to connect to the given URL just like a regular user so it's a more realistic test of whether the site is available.

Quote:

Originally Posted by Turbocapitalist

Or if you are getting into more serious numbers of servers and services then look at tools like Zabbix

If you're building an internal monitoring system then also consider Nagios or the fork Centreon.

At the moment I'm using a central Centreon instance with remote pollers to monitor 3500 service on 350 hosts across three data centers, the pollers also monitor each other and can send alerts independently of the central instance. There's also external with NodePing for 3rd party remote checking of connectivity to sites. Externally we use Site24x7 to check specific URLs, and I've just noticed that Site24x7 have a free plan with e-mail alerting for up to 5 URLs

Wayne Sallee · 01-19-2019, 08:54 AM

You could set up a server in another location to do the checks. Any linux computer in a different location, can be easily set up to do this for you.

Wayne Sallee
Wayne@WayneSallee.com
http://www.WayneSallee.com

seanmancini · 01-22-2019, 08:25 AM

I use https://uptimerobot.com/

They have a free basic web check based on HTTP responses it has worked well for me

Sean Mancini
https://www.seanmancini.com
sean@seanmancini.com

rnturn · 09-21-2019, 06:30 AM

Quote:

Originally Posted by TenTenths

At the moment I'm using a central Centreon instance with remote pollers to monitor 3500 service on 350 hosts across three data centers, the pollers also monitor each other and can send alerts independently of the central instance. There's also external with NodePing for 3rd party remote checking of connectivity to sites. Externally we use Site24x7 to check specific URLs, and I've just noticed that Site24x7 have a free plan with e-mail alerting for up to 5 URLs

When I was wrangling a Nagios monitoring system, we made sure to also monitor the routers as well as the hosts. That way we could keep the alerts due to a router outage down to just the router instead alerts for everything that Nagios couldn't reach through that router. The server team was especially happy to not receive alerts for something that the network team needed to handle.

Cheers...

gani · 09-21-2019, 08:41 AM

You can try what I have done to automatically monitor via cron crucial processes that are just suddenly shutting down with no apparent reason. The script will verify the process PID, if not available, it means the process has shutdown and the script will start it on its own. This test script was done on my Linuxmint (Ubuntu) desktop, just do necessary adjustments on your side. Here is the script:

Code:

PID=$(pidof apache2)
if [ "$PID" = "" ]; then
    rm -f /var/run/apache2/*.pid
    systemctl start apache2.service # or /etc/init.d/apache2 start
    if [ "$(pidof apache2)" != "" ]; then
        echo "apache2 restarted from unexpected shutdown..."
    fi	
fi

You may place this script in /root or in your choice. Say name it as: chkapache2.sh. You may first test the script by shutting down apache2 first.

Crontab entries (Just my periodic schedule suggestion):

59 22 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 3 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 5 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 7 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 10 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 12 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 15 * * * sh /root/chkapache2.sh > /dev/null 2>&1
0 18 * * * sh /root/chkapache2.sh > /dev/null 2>&1

I've been doing this on my running mailserver on Slackware that amavisd-new, postfix and even postgrey is just suddenly shutting down after some long period of operations.

TenTenths · 09-23-2019, 04:07 AM

Quote:

Originally Posted by rnturn

When I was wrangling a Nagios monitoring system, we made sure to also monitor the routers as well as the hosts. That way we could keep the alerts due to a router outage down to just the router instead alerts for everything that Nagios couldn't reach through that router. The server team was especially happy to not receive alerts for something that the network team needed to handle.

Cheers...

I used hosts as a generic term, in our situation this includes routers, switches, UPS, etc. In my previous role it also included environmental sensors such as temperature, humidity, moisture sensors, etc. It's NOT fun to get an SMS message that the water detection sensor in the comms room floor has activated

tyler2016 · 09-24-2019, 09:26 AM

I use Zabbix to monitor my stuff. I am pretty happy with it. It does have a little bit of a learning curve, but I picked it up pretty quickly.

bathory · 09-24-2019, 01:36 PM

Monit