Hello, it would be nice if monitorix has the capability to monitor the cpu usage of processes, like 'top' does. It could display the 5-processes with the most cpu usage in a nice looking graph.

I had an unfortunate system crash on a host that is using monitorix and I have noticed that it won't start anymore. I am using the vanilla docs/monitorix.init boot script and I think it has a bug which causes the startup problem in my cause.

The start() function in this file makes sure that there's no lock file and then starts monitorix.

But the opposite use case (a present lock file) can happen in more than one scenario:

1. monitorix is already running, everything is fine, there's no need to start it again
2. my scenario: system has crashed and was restarted. monitorix is not running, yet there's the stale lock file, the boot script sees it and never starts monitorix

Conclusion: it is not enough to check for a present lock file in start() function. The boot script also has to explicitly check if monitorix is already running. If monitorix is not running, the boot script should assume that the lock file is stale, remove it and then start monitorix the normal way.

Hi,

HTTP built-in server is great but missing one "must be" feature:

• access only for username/password
• access only for IP range

For now I switch back to Apache but it will be great use only built-in server from Monitorix.

best regards

I set up two file system groups to monitor. I then wanted to monitor the FS usage for my second group in a similar manner to how the multi-host web page is put together, i.e. using iframes. However, the results produced by the expected urls are not consistent. For instance something like

http://myhost:88/monitorix-cgi/monitorix.cgi?mode=localhost&when=1day&graph=_fs1&color=black&silent=imagetag

will produce the expected FS usage graph for the first group. However, something like

http://myhost:88/monitorix-cgi/monitorix.cgi?mode=localhost&when=1day&graph=_fs2&color=black&silent=imagetag

produces a page with the expected disk I/O activity for the first group, but also the filesystem usage for my second group (see attached image). The group numbers here are 0 and 1.

How difficult would it be to install Monitorix without its built-in webserver?

Hello,

The "port" feature make graph with iptables (monitorix_*) (ipv4), the ipv6 (ipv6tables) graph not support.

Could You add this feature ?

Thanks
Tarjaizaid

Jodri - How can I modify monitorix to allow plots of the network graph to be on the log scale? Sometimes, my network graph gets this enormous spike it in as you can see below. Plotting on the log scale would help to even this out.

If you (like me) have a lot of disks, you need to break the graphs into multiple ones in order to show statistics for all of them. But in email reports, only statistics for the first group is included.
It would be great if the email reports included statistics for all fs/disk groups.

Thanks!
Best regards, Patrick

udev is a strange beast that sometimes may decide to reorder devices for no apparent reason, resulting in /dev/sda being the device that until yesterday was /dev/sdb...

When this happens monitorix continues to plot temperatures without realising that drive "letter" allocation has changed.

The best way of accepting this is to fully support the use of "/dev/disk/by-path/" paths under

Hi!

In my infrastructure, I use hiawatha webserver. An easy way to open monitorix to the wild internet is to use the reverse proxy possibility of hiawatha. This way, I just have to configure monitorix to listen to localhost connection through its builtin http server. It seems to works out of the box, but.... No graphs are displayed (all images are broken). This comes from the fact you force $config{url} to be a combination of hostname and port config value.

I suggest to do the following modification to allow a new url config value, which will take precedence over the hostname/port default if present.

--- /srv/http/monitorix/cgi/monitorix.cgi   2014-08-09 18:13:34.000000000 +0200
+++ monitorix.cgi   2014-08-10 13:16:39.252252098 +0200
@@ -257,11 +257,13 @@
    }
 }

-$config{url} = ($ENV{HTTPS} || ($config{httpd_builtin}->{https_url} || "n") eq "y") ? "https://" . $ENV{HTTP_HOST} : "http://" . $ENV{HTTP_HOST};
-$config{hostname} = $config{hostname} || $ENV{SERVER_NAME};
-if(!($config{hostname})) { # called from the command line
-   $config{hostname} = "127.0.0.1";
-   $config{url} = "http://127.0.0.1";
+if(!$config{url}){
+   $config{url} = ($ENV{HTTPS} || ($config{httpd_builtin}->{https_url} || "n") eq "y") ? "https://" . $ENV{HTTP_HOST} : "http://" . $ENV{HTTP_HOST};
+   $config{hostname} = $config{hostname} || $ENV{SERVER_NAME};
+   if(!($config{hostname})) {  # called from the command line
+       $config{hostname} = "127.0.0.1";
+       $config{url} = "http://127.0.0.1";
+   }
 }
 $config{url} .= $config{base_url};

Good morning,

Finally we could update our Monitorix servers to 3.5.1, and looking deeply on monitorix.conf, we found the Email reports feature (We where on 2.x release and enabled only the needed graphs).
So, I can write the smtp hostname and from fields, but where does the password can be typed? As well, does it support TLS, with self-signed certificates?
Hoping to hear good news or be implemented in a next release,

Greetings

The hostname is set to my server's hostname in /etc/monitorix.conf but neither the front page nor the details page shows it. Instead, it shows "localhost" in both.

Hello,

when i use process feature and enable show_gaps option, the web page generate an error for "Opened Files" graph.
"ERROR: while graphing /var/lib/monitorix/www//imgs/process05z.1week.png: parameter 'wrongdata' does not represent a number in line AREA:wrongdata#FFFFFF:"

If i disable show_gaps, i dont have this error.

Thx

In my fs graph, I have listed "/, [a RAMdisk], /swapfile"

The /swapfile usage graph line tracks to the RAMdisk's actual (correct) usage, though "free" shows that of 2.1G of swap, 0B are used.

Hi,

I followed the basic instructions and installed RPM from rpm -ivh http://www.monitorix.org/monitorix-n.n.n-1.noarch.rpm

As I am going through my /etc/monitorix.conf and comparing to http://www.monitorix.org/documentation.html#33

I see you dont have the monthly section defined in the sample conf. I also checked the master repo conf file and its also missing.

Am I missing something?

Thanks for your great work.

Hello,
I'm getting following warnings in monitorix.log:

Use of uninitialized value $cachf in subtraction (-) at /usr/lib/monitorix/phpapc.pm line 169.
Use of uninitialized value $cachf in numeric ne (!=) at /usr/lib/monitorix/phpapc.pm line 170.
Use of uninitialized value $cache_suffix in string eq at /usr/lib/monitorix/phpapc.pm line 175.
Use of uninitialized value $cache_suffix in string eq at /usr/lib/monitorix/phpapc.pm line 176.
Use of uninitialized value $cachf in concatenation (.) or string at /usr/lib/monitorix/phpapc.pm line 180.
Use of uninitialized value $cachs in concatenation (.) or string at /usr/lib/monitorix/phpapc.pm line 180.

The php-apc graph is displaying correctly.

Monitorix: 3.4.0
perl: 5.18.1

Regards

Currently, there appears to be no way of specifying an interface (address or hostname) for the standalone web server to bind to. This leaves the standalone web server open to listening across all interfaces. While, in theory, it is not that big of a deal, it would be nice to have the ability to refine the listening scope of the standalone web server.

Hello

I have setup everything and got my Monitorix install to work as I want.
I also have a working installation of Fail2Ban which is working fine. It is blocking and logging to the required log file.
I cannot get Monitorix to graph Fail2Ban at all.
I have plenty of entries in the log file.
Defined jails correctly.
But Monitorix shows only 2 SSH jails max and no entries for the other jails

monitorix-3.5.1

Old version of proprietary NVidia driver: 331.67

Everything worked well with this one

After upgrading to 340.24 the NVidia temperature graph keeps staying empty

Both nvidia-settings and nvidia-smi show the correct value of the GPU temperature

Nevermind... false alarm.

Hi! I threw monitorix on my server a couple weeks ago, and just noticed that the disk module is behaving inconsistently with the FAQ page. It states that I can use my RAID controller and its respective disks if it is supported by smartctl by supplying the appropriate argument for smartctl along with my controller's device. However, my monitorix log file reports that the device is invalid. It's running with a CCISS controller, which is the very example shown on the FAQ page that I was modeling with my config.

Log Entries Produced on Monitorix Start

Sun Apr 28 12:21:27 2013 - Starting Monitorix version 3.1.0 (pid 19668).
Sun Apr 28 12:21:27 2013 - ERROR: disk::disk_init: invalid or inexistent device name '/dev/cciss/c0d0 -d cciss,0'.

Disk Module Config

<disk>
        <list>
                0 = "/dev/cciss/c0d0 -d cciss,0", "/dev/cciss/c0d0 -d cciss,1", "/dev/cciss/c0d0 -d cciss,2", "/dev/cciss/c0d0 -d cciss,3", "/dev/cciss/c0d0 -d cciss,4", "/dev/cciss/c0d0 -d cciss,5"
       </list>
</disk>

Smartctl Working Perfectly Fine, Indicating some Issue with Monitorix

[root@WuffServer /etc]# for i in $(seq 0 5); do smartctl /dev/cciss/c0d0 -d cciss,${i} -a | grep 'Current Drive Temperature'; done 
Current Drive Temperature:     20 C
Current Drive Temperature:     20 C
Current Drive Temperature:     20 C
Current Drive Temperature:     20 C
Current Drive Temperature:     20 C
Current Drive Temperature:     19 C

System Information

OS: Arch Linux
Monitorix Version: 3.1.0-8
Smartmontools Version: 6.1-2
Hardware: Proliant DL380 G3

Example of Graph Output

Note, these should not be a huge issue if utilizing a proper firewall. But it might be something you would want to look into and fix anyways.

These where found by a Nessus scan performed against a server running Monitorix.

First One:

Severity: HIGH

Exploit:

Goscript go.cgi Arbitrary Command Execution

Description

The remote host is running GoScript. The installed version fails to properly sanitize user-supplied input to the 'go.cgi' script. An unauthenticated, remote attacker could exploit this flaw to execute arbitrary commands on the remote host.

Solution

There is no known solution at this time.

See Also

http://archives.neohapsis.com/archives/bugtraq/2004-08/0037.html

Plugin Output

It was possible to execute the command 'id' on the remote host
by requesting the following URL :

http://10.19.75.76:8080/amPortal/action/go.cgi|id|

Second One:

Severity: Medium

Exploit

Web Server Generic Cookie Injection

Description

The remote host is running a web server that fails to adequately sanitize request strings of malicious JavaScript. By leveraging this issue, an attacker may be able to inject arbitrary cookies. Depending on the structure of the web application, it may be possible to launch a 'session fixation' attack using this mechanism.

Solution

Contact the vendor for a patch or upgrade.

See Also

http://en.wikipedia.org/wiki/Session_fixation
http://www.owasp.org/index.php/Session_Fixation
http://www.acros.si/papers/session_fixation.pdf
http://projects.webappsec.org/Session-Fixation

Plugin Output

The request string used to detect this flaw was :

/<script>document.cookie=%22testzgsf=9268;%22</script>

The output was :

HTTP/1.0 404 Not found
Date: Wed, 20 Nov 2013 13:34:54 -0500
Server: Monitorix HTTP Server
Connection: close
Content-Type: text/html; charset=UTF-8

[...]

Not Found

The requested URL /<script>document.cookie="testzgsf=9268;"</script> was not found on this server.

Third one:

Severity: Medium

Exploit

Web Server Generic XSS

Description

The remote host is running a web server that fails to adequately sanitize request strings of malicious JavaScript. By leveraging this issue, an attacker may be able to cause arbitrary HTML and script code to be executed in a user's browser within the security context of the affected site.

Solution

Contact the vendor for a patch or upgrade.

See Also

http://en.wikipedia.org/wiki/Cross-site_scripting

Plugin Output

The request string used to detect this flaw was :

/<script>cross_site_scripting.nasl</script>.asp

The output was :

HTTP/1.0 404 Not found
Date: Wed, 20 Nov 2013 13:36:20 -0500
Server: Monitorix HTTP Server
Connection: close
Content-Type: text/html; charset=UTF-8

Not Found

The requested URL /<script>cross_site_scripting.nasl</script>.asp was no
t found on this server.

Hi,

i have a suggestion...
you can add a timestap on graphs to recorded the actuality of the graphs

greets pr0vieh

This is a snapshot of the net traffic graph for my Raspberry Pi:

The peak is because 4 days ago I transferred some big files from the Raspberry Pi to my laptop via local network. I'm not interested in such values, so I tried to zoom on smaller values (to crop-out the peak) setting the rigid/limit in the monitorix.conf:

<net>
        list = eth0
        <desc>
                eth0 = Ethernet LAN, 600, 600
        </desc>
        gateway = eth0
</net>

I tried also with 0, 600 and 600, 0, but the graph is not changing. Am I missing something, will the limits be applied only for values collected from now (no retroactive effect), or is this a bug?

Using the the new config file:

Starting monitorix: Config::General: Cannot create hashref from because there is already a scalar option 'theme' with value 'black' at /usr/bin/monitorix line 363.

False alarm. Not a problem with your code. Fixed. Plz close.

In my monitorix.conf i wrote:
conn_type = socket
list = /var/run/mysqld/mysqld.sock

/var/run/mysqld/mysqld.sock = 3306, my_mysql_user, my_mysql_password

But Monitorix all the same use user 'user' with password 'secret' for authorization on MySQL

Hi there,
any graph for memcache ?
like : http://livebookmark.net/journal/2008/05/21/memcachephp-stats-like-apcphp/

thanks

Why nocompress in logrotate?

Hi Jordi,

Any chance of getting some MongoDB monitoring support added to Monitorix?

More then willing to be a test dummy :-)

Hi Jordi

we ran in to a snag today. It seems Monitorix is dependent on the mysql-libs from 5.1

Our problem arises from a conflict when running a newer version of MySQl. Is there anyway to easily bypass this issue. would Installing the application with no dependencies cause an issue by chance?

Installing:
monitorix noarch 3.5.1-1.el6 epel 173 k
Installing for dependencies:
mysql-libs x86_64 5.1.73-3.el6_5 updates 1.2 M
perl-DBD-MySQL x86_64 4.013-3.el6 base 134 k

Transaction Check Error:
file /usr/share/mysql/charsets/Index.xml from install of mysql-libs-5.1.73-3.el6_5.x86_64 conflicts with file from package MySQL-server-5.6.20-1.el6.x86_64

I had troubles using Monitorix (installed via IzzySoft repo) with Apache web server. Using the given Apache configuration file (the one in /usr/share/doc/monitorix/monitorix-apache.conf) I had _403 Forbidden_ all the time. I wrote down all the details in this Server Fault question.

The solution came from this answer. I had to run sudo a2enmod cgi and edit the configuration file this way:

Alias /monitorix /var/lib/monitorix/www
ScriptAlias /monitorix-cgi /var/lib/monitorix/www/cgi

<Directory /var/lib/monitorix/www/>
    <IfModule mod_authz_core.c>
        # Apache 2.4
        Require all granted
    </IfModule>
    <IfModule !mod_authz_core.c>
        # Apache 2.2
        Order deny,allow
        Allow from all
    </IfModule>
</Directory>

<Directory /var/lib/monitorix/www/cgi/>
        DirectoryIndex monitorix.cgi
        Options ExecCGI
</Directory>

Unfortunately I can't tell you if if this problem affects only Apache 2.4, only Ubuntu 14.04, or only both combined.

Hi,
today I have set up monitorix instead of munin. It is working very well, but the postfix graphs do not seem to work. I have installed syslog-ng to ensure the needed logfiles and pflogsumm as mentioned in the docs.

I have set my log paths to this:

# Log files pathnames
# -----------------------------------------------------------------------------
log_file        = /var/log/monitorix
secure_log      = /var/log/auth.log
mail_log        = /var/log/mail.log
milter_gl       = /var/milter-greylist/greylist.db
imap_log        = /var/log/mail.log
hylafax_log     = /var/spool/hylafax/etc/xferfaxlog
cups_log        = /var/log/cups/page_log
ftp_log         = /var/log/vsftpd.log
fail2ban_log    = /var/log/fail2ban.log
spamassassin_log        = /var/log/mail.log
clamav_log      = /var/log/clamav/clamd.log
cg_logdir       = /var/CommuniGate/SystemLogs/
squid_log       = /var/log/squid/access.log

The graphs only show the queued messages. Any ideas? I am using arch linux.

Stefan

Hey,

I would like to embed the graphs on another website, I thought it would be enough by just putting in the link and monitorix auto refresh them every x minutes. But it doesn´t seem like it. So is there any option or something I could run as a cronjob to refresh the picture graphs, that I don´t have to open the monitorix website ?

Hello,
There is a cosmetic bug in gaps with no data display - gaps are displayed only for positive values. For negative values no gap is displayed (for example in mail statistics):

Using monitorix 3.1.0 ubuntu 12.04

When using any interface device other than eth0, my network usage graphs indicate that I get petabyte/s transfer speeds. I checked, and I'm not downloading the internet every few seconds, so some averaging must be off. Here is one example config and /proc/net/dev output. I'd be happy to dig up more info if needed.

========= Config File Snip =====================

Inter-| Receive | Transmit
face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
bond0: 163823767473 135997226 0 830762 0 0 0 216546 98954858466 102127603 0 0 0 0 0 0
eth0: 37094012749 36656010 0 773966 0 0 0 108085 49352643953 47646487 0 0 0 0 0 0
eth1: 126729754724 99341216 0 0 0 0 0 108461 49602214513 54481116 0 0 0 0 0 0
lo: 10017646 128240 0 0 0 0 0 0 10017646 128240 0 0 0 0 0 0

After upgrde to 3.5.1 MySQL stats stopped working:

Mon May 19 10:01:01 2014 - mysql::mysql_update: Cannot connect to MySQL '/var/run/mysqld/mysqld.sock'.
Mon May 19 10:01:01 2014 - ERROR: while updating /var/lib/monitorix/mysql.rrd: /var/lib/monitorix/mysql.rrd: expected timestamp not found in data source from N

MySQL listens on socket - tested with socat.

Is it possible for you to include support for the Raspberry Pi in the existing CPU temp and voltage plots?

The raspberry pi firmware comes with a util to generate these data which is by default installed to: /opt/vc/bin/vcgencmd

CPU

% vcgencmd measure_temp
temp=48.5'C

CPU Voltage

% vcgencmd measure_volts core
volt=1.20V

Thanks for the consideration!

Hello,

Great project ! I'm using it from few years ago ... on all my servers.

There is a minor bug in /usr/bin/monitorix - if you take mysql statistics and connecting via socket, it returns:

DBI connect('mysql_socket=/var/run/mysqld/mysqld.sock','HASH(0x2b5d0d8)',...) failed: Access denied for user 'HASH(0x2b5d0d8)'@'localhost' (using password: NO) at /usr/bin/monitorix line 4475

That line it's:

$dbh = DBI->connect( "DBI:mysql:mysql_socket=$sock", { PrintError => $print_error, } ) or logger("$myself: Cannot connect to MySQL '$sock'.") and next;

it should be:

$dbh = DBI->connect( "DBI:mysql:mysql_socket=$sock", $user, $pass, { PrintError => $print_error, } ) or logger("$myself: Cannot connect to MySQL '$sock'.") and next;

Modified in my script and works flawlessly.
Have a good day !

75e7b93 is an important fix. I have been using the git version for a week now and it has been very stable. Please consider wrapping up the fixes into a formal version. Thanks!

The temperature sensor of my ati graphics card can be displayed by lm_sensors. Unfortunately, monitorix doesn't allow to set the gpu-keys to anything different than nvidia.
For an older version I quickly hacked something together by basically copying the logic for the cpu-keys and adopting it for the gpu, which worked perfectly.
It would be nice, if that would be a feature that is in monitorix by default.

Hello,

I think I have a problem with ipv6 netstat statistics.
The graph show correct "listen connections" but wrong value for "established connections".
I count 5 "established connections" for IPv6.

What did you think ?

Thanks
Tarjaizaid

Seems as though the nvidia binary blob again supplies the ability to measure the GPU temp which was removed from previous versions.

Output on an old GeForce 6200:

% nvidia-smi -q -d TEMPERATURE

==============NVSMI LOG==============

Timestamp                       : Thu Oct 10 08:12:46 2013
Driver Version                  : 304.108

Attached GPUs                   : 1
GPU 0000:01:05.0
    Temperature
        Gpu                     : 49 C

Output from an embedded GeForce 9400M:

% nvidia-smi -q -d TEMPERATURE

==============NVSMI LOG==============

Timestamp                           : Thu Oct 10 08:12:21 2013
Driver Version                      : 325.15

Attached GPUs                       : 1
GPU 0000:03:00.0
    Temperature
        Gpu                         : 39 C

There appears to be no way to enable port logging of both TCP and UDP connections on the same port, as you might want to do for logging bittorrent-like applications.
Tried:
(1) using 'tcp/udp' as option
(2) separating entries into a 'tcp' line and a 'udp' line
(3) (1) and (2) with desired port in enable list twice
to no avail

When viewing http://localhost/monitorix, the pull down that lists various hosts should contain the hostname rather than localhost, shouldn't it?

What about using HTML templates to generate the Monitorix initial page (the one with the host/graph/time form) and the graphs page?

I wanted to customize the styling a bit (table thick borders, big fonts, etc), but I noticed all HTML generation is hard-coded in Perl scripts. Unfortunately I'm a PHP developer, and I know nothing about Perl. If I was, not only I had modified my pages, but also helped developing a template system here.

After 15 min. of working stats from BIND, monitorix stop make graphs.

Fri Mar 29 17:59:00 2013 - main::ALRM_handler: bind_update(): Can't use string ("ANY") as a HASH ref while "strict refs" in use at /usr/lib/monitorix/bind.pm line 273.

Fri Mar 29 18:00:01 2013 - main::ALRM_handler: bind_update(): Can't use string ("ANY") as a HASH ref while "strict refs" in use at /usr/lib/monitorix/bind.pm line 273.

Hello,
The mix/max/avg values for mail statistics are displayed incorrectly. Here is an example of one month:

I have try to

1. add (into the monitorix.cgi)
   print("<link rel='stylesheet' type='text/css' href='style.css'>\n");
   at line 382 under print("<title>$config{title}</title>\n");
2. and copy the style.css into all the dirs of the path /var/lib/monitorix/www/cgi (centos 6)

but allays i get the error NOTEXIST - [192.168.1.68] File does not exist: /monitorix-cgi/style.css

also i have try other paths at step 1 like:
href='style.css'
href='/style.css'
href='/var/lib/monitorix/www/cgi/style.css'
href='/var/lib/monitorix/www/style.css'
href='/var/lib/monitorix/style.css'
href='/var/lib/style.css'
href='/var/style.css'
href='/style.css'

but always i get the NOTEXIST error in the log.

any ideas how to add a style.css file for theming monitorix?

thanks for this very useful application!

With 3.0, it seems the order I list my mount points in the fs section change the mapping made by Monitorix.

The fs section look like this:

<fs>
        <list>
                0 = /, /home, /boot, swap, /data
        </list>
        <desc>
        </desc>
        <devmap>
        </devmap>
        rigid = 2, 0, 0
        limit = 100, 1000, 1000
        <alerts>
                rootfs_enabled = n
                rootfs_timeintvl = 3600
                rootfs_threshold = 100
                rootfs_script = /path/to/script.sh
        </alerts>
</fs>

df output looks like this:

Filesystem      1K-blocks       Used Available Use% Mounted on
rootfs          100798068   66993476  28684276  71% /
dev               1944508          0   1944508   0% /dev
run               1947480        712   1946768   1% /run
/dev/sda3       100798068   66993476  28684276  71% /
tmpfs             1947480      60152   1887328   4% /dev/shm
tmpfs             1947480          0   1947480   0% /sys/fs/cgroup
tmpfs             1947480        468   1947012   1% /tmp
/dev/sda1          101086      40821     55046  43% /boot
/dev/sdb3      1922729784 1069397472 755663248  59% /data
/dev/sda4       379557808   10646140 349631260   3% /home

Below is the resulting graph with 2.6.0, then I switched to 3.0 with 3 different list orders:

0 = /, swap, /boot, /home, /data
0 = /data, /home, /, swap, /boot
0 = /, /home, /boot, swap, /data