Thold 2.x - Thresholding/Alerting module for cacti 8.6

alexus · Post by **alexus** » Wed Oct 05, 2005 8:09 am

so which one is the latest

thold2.1b-0.8f.tar.gz or thold2.1a-0.8g?

i'm really curiose who's author of that script/pluging.. why cant we make this script to go to official and make it more stable, it seems like a lot of people having issues with that..

AndyStewart · Post by **AndyStewart** » Thu Oct 06, 2005 4:46 am

Since installing THOLD, i've found that viewing any of the graphs, or trying to modify any devices is now INCREDIBLY slow. Whereas before pages would show up within 2-3 seconds, its now taking over 60 seconds (and sometimes longer) before the pages finally refresh...

Any ideas?

Just a test server at the moment, but still a reasonable spec (3.2 GHZ proc, 256 MB Mem, 40 GB disk space, running 2003 fully patched apart from SP1. All CACTI info in signature)

Have had to disable the THOLD module for now (and speed has returned to normal - ie very quick) but i really want to use this module...

AndyStewart · Post by **AndyStewart** » Mon Oct 10, 2005 3:59 am

/bump.... still got thold disabled due to it grinding machine to a halt - anyone seen this before or got any recommendations?

AndyStewart · Post by **AndyStewart** » Tue Oct 11, 2005 3:35 am

thold now enabled. Found out what was killing the page...

i've commented out the following line from thold.php (line 131 or thereabouts) and now everything is working fine:

<img src="<?php echo $config["url_path"]; ?>graph_image.php?local_graph_id=<?php echo $graph ?>&rra_id=1&graph_start=-32400&graph_height=100&graph_width=300&graph_nolegend=true">

If i leave this code in, then trying to set up a threshold just hangs the machine on this page until you click stop and then the page appears without the graph. I then have to stop/restart the WWW service to get the site working again.

eddievenus · Post by **eddievenus** » Fri Oct 14, 2005 1:46 pm

I have looked in the forums for the solution to this issue, but come up short each time. Perhaps someone here can help me find where my setup went wrong.

I am getting timestamps on emails, sent by thold via my mailserver (running surgemail), that is an hour in the future on all Host down messages.

I did not notice this before, but cannot say that it was not happening, since I am rarely in the office when these messages come in and thus do not often get to corrolate the message with the graphs.

However for the last few days I have been actively monitoring a few key devices, and noticed that the graphs show outages at say 11:45am. I will get a message at 11:46am telling me it is down, but the timestamp on the email is 12:45. Then it will come back online at 12:10pm, I will get my Host returned message, a minute or 2 later, but the time stamp is 1:10, not 12:10.

I looked into this, the mailserver is the Cacti server (and the thold server). The mail app is Surgemail, which I have used for years without issues. The clock is set for the right time, the timezone is correct and it is auto updating the clock every day, as instructed.

I have tested email to and from various applications, both on the mail server directly and from other network devices sending through the mail server. All those timestamps are correct.

How can the timestamp for this one application (thold or php) be off by exactly 1 hour? If other apps were affected I could guess that it was a mail server issue, but they are not. I can send mail from the server to other mail accounts, including the one I use for monitoring, and always get the right timestamp. So I have to presume that Thold is the only variable that makes a difference in this equasion.

I would appreciate even the least bit of help with this as it is frusterating to look back on old email from thold that I have presumed to have accurate timestamps on, and relize that that presumtion may have been incorrect.

thanks
eddievenus

p.s. system specs in case it matters
win xp pro sp2
IIS 5.1
cacti 0.8.6f
thold 2.1a (with suggested fixes for known issues)
php poller
1GB RAM, 1+ TB HD
althon xp 2400+
net-snmp 5.x
php 5
rrdtool 1.2x

dforsdike · Post by **dforsdike** » Wed Oct 19, 2005 5:07 am

Hi

I have installed the threshold addon and everything seems to be fine except for the current value is always 0, so the monitors are never triggered.

How should the check-thold.php scheduled task be setup, should it be set to run at the same time as the cacti poller or different time and how often? i currently have it set to every minute.

I have seen a post earlier on regarding this issue by changing code in the get_current_value functio, but still this didnt resolve my issue.

Can anybody help?

I am running:
Windows 2003 Sp1
PHP 5.0.2
Mysql 4.0.21
Cacti 0.8.6G

dforsdike · Post by **dforsdike** » Wed Oct 19, 2005 5:07 am

Hi

I have installed the threshold addon and everything seems to be fine except for the current value is always 0, so the monitors are never triggered.

How should the check-thold.php scheduled task be setup, should it be set to run at the same time as the cacti poller or different time and how often? i currently have it set to every minute.

I have seen a post earlier on regarding this issue by changing code in the get_current_value functio, but still this didnt resolve my issue.

Can anybody help?

I am running:
Windows 2003 Sp1
PHP 5.0.2
Mysql 4.0.21
Cacti 0.8.6G

sa · Post by sa » Sun Oct 23, 2005 7:34 pm

i'm a little bit confused. i installed thold 2.1a on my cact 0.8.6g patched version, and all seems to work.

exactly, i'm confused on what thold plugin intend for "current value"

on my graphs i see 420 kbps out and 84 kbps in, on threshold i see 51630 out and 9934 in.

someone can explain me that?

tnx

eddievenus · Post by **eddievenus** » Mon Oct 24, 2005 9:24 pm

dforsdike

I think I can help. I ran into this same exact issue, and thanks to the post you are talking about I was able to resolve it.

I am also checking every minute, but I am also using a little app I found called Visual Cron (for windows) which is nice when you want to set up tasks since it can do alot more than task scheduler. However the part that sparked my interest is the part that allows you to run jobs at a specific second of every minute. So I have the poller script running every minute on the minute, and the thold script running every minute 45 seconds later. This leaves plenty of time, since my poller takes betweeen 9 and 40 seconds depending on conditions of the internet, my LAN, its usage, etc. So this little tool allowed me to get around the timing issue until I tried the code fix.

Once I made the change to the code however it worked like a charm, in fact it was suggested to set it to 900, I changed that after reading some other posts in other threads, and set mine to 60 or something like that.

Good luck, I hope that helps.

Ehud · Post by **Ehud** » Tue Oct 25, 2005 7:03 am

I have fedora core 4 and cacti 0.8g , how can I install it ?

Ehud

N3NCY · Post by **N3NCY** » Tue Oct 25, 2005 12:31 pm

I cant seem to get data in the "Current" field on the new Threshold tab.
Hosts say they are UP (Green) and my graph shows data.
Example, my Cisco switch running CPU at 13% according my graph.
I set the high threshold at 15 and make the switch work harder.
The graph shows the CPU going to 17% but no alert.
The "Current" field never shows any data values.

Is my problem "G" 0.8.6g?

UPDATE!!!! Doh!
ANSWER:
I had to specify the full path to PHP in my crontab!

*/5 * * * * root /usr/local/bin/php /opt/apache/htdocs/cacti/poller.php >/dev/null 2>&1 && /usr/local/bin/php /opt/apache/htdo
cs/cacti/thold/check-thold.php > /dev/null 2>&1

Markus21 · Post by **Markus21** » Wed Oct 26, 2005 12:30 am

so my englisch is not so god, sorry,

i have installed the plugin "thold 2.5" and i let run cacti under windows 2003 server and xp professionell, so i can´t send a test email under the tab thold/alerting under settings in cacti. with mozilla, opera and so .. it goes fine, but not under internet explorer (IE), my firm is using windows and internet explorer, it must go with him, i hope you are using internet explorer and you know the problem,

thanks for your help in advance

gruss markus21

mvam · Post by **mvam** » Wed Oct 26, 2005 4:10 pm

i just installed the plugin arch and thold 2.5 and setup some thresholds. i set the values to 60 high and 5 low. when the poller runs the threshold module thinks everything is below 5 and is constantly sending email. i decided to run check-thold.php manually and this was the result:

10/26/2005 01:58:55 PM - POLLER: Poller[0] Checking Thresholds
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?
ERROR: end time: did you really mean month 899?

it sends another email at the 12th cycle, but none seem to go into a "not triggered" state.

Markus21 · Post by **Markus21** » Thu Oct 27, 2005 12:19 am

the site http://cactiusers.org, is not rechable since 3 - 4 days, everbody now,
what happend, or chanced the url ?

i want to download the plugin thold, monitor and report,

can i download it from another site, when yes where ?

psyber · Post by **psyber** » Thu Oct 27, 2005 1:39 am

odd i get there just fine (btw your url has a coma in it at the end)

http://cactiusers.org/index.php

Cacti

Thold 2.x - Thresholding/Alerting module for cacti 8.6

How do I install it onn LInux ?

Thold version for Cacti 0.8.6g?

Thold Plugin with Internet Explorer / IE , help !!!!!!!

morning,

Who is online