Cactiusers.org

I have the latest version of thold running on cacti 0.8.6g on a Windows XP machine. Everything is working great except the notifications. It seems that I only get notifications when I have baseline monitoring enabled and baseline notifications enabled. Is it true that baseline must be enabled for high and low threshold triggers to work? My problem is with high and low thresholds. If I set a high threshold and my traffic is high enough to trigger it, it is listed as triggered on the thresholds page but I'm not getting an e-mail notification everytime. I have the trigger set at 1. What is strange is that about 25% of the time I get a notification like I should when the low threshold is triggered but the rest of the time the high or low threshold is triggered but no e-mail is sent. However, with baseline monitoring anytime baselines trigger high or low then I get an e-mail like I should. Any ideas??

One last question, what are the suggested settings for Baseline Reference, Baseline time range and Basline Deviation when monitoring bandwidth use? I'm currently using 86400, 10800 and 25%.

Its strange that it would email only 25% of the time. Its usually an everytime or never issue. I really need to beef up the debugging capabilities of plugin, to log everytime it sends an email. You can currently run check-thold.php with the command line switches
/show /debug
to see what is happening, but it doesn't specify when it sends an email or not. But if the sending happens to error out, it will display the reason why. You might want to test with a easier to trigger threshold to see if you can reproduce it reliably, if so then we may be able to pinpoint the problem.

It actually seems that it only sends mail (and very rarely in this case) when it drops below the low threshold and never sends it when it goes above. I ran /show /debug and this is what I get for the one that is dropping below the threshold but I'm not getting an e-mail:

C:\Inetpub\wwwroot\cacti\plugins\thold>php check-thold.php /show /debug
12/29/2005 10:26:26 AM - THOLD: Poller[0] Checking Thresholds
Checking Threshold : "Switch 1 - Traffic - Client 1"
Data Source : traffic_out is normal with 131804
RRA: 49 : traffic_out
Ref. values count: 8
Ref. value (min): 83269
Ref. value (max): 243237
Cur. value: 131804
Low bl thresh: 41635
High bl thresh: 316208
Check against baseline: OK
------------------
Checking Threshold : "Switch 1 - Traffic - Client 1"
Data Source : traffic_in went below threshold of 50811 with 24717
Checking Threshold : "Switch 1 - Traffic - Client 2"
Data Source : traffic_in is normal with 5505
RRA: 48 : traffic_in
Ref. values count: 8
Ref. value (min): 11327
Ref. value (max): 131857
Cur. value: 5505
Low bl thresh: 1133
High bl thresh: 184600
Check against baseline: OK
------------------

It looks like when the traffic_in went below the threshold value it just stops doing anymore checking of that port. It cuts off the log entry immediately and goes to the next port. Do you see anything that helps?

I took off baseline checking. I now only have low and high triggers. After several days of testing on my live cacti server the low threshold is working properly. Everytime one of my hosts goes below the low threshold I get an e-mail. However, I have never received an e-mail when the high threshold is triggered. I just had 3 different hosts trigger a high threshold and I received no e-mails for any of them. Any idea why this would be happening?

I'm still only getting low threshold e-mails. I've looked at everything and can't seem to figure out why the high thresholds are not sending. Everythign else works fine. High thresholds are triggered correctly but I never get an e-mail for them. Please any help with this would be very appreciated.

when you say you turned off base line monitoring did you clear out all the boxes or did you just uncheck the baseline monitoring box?


Not that I think it should matter, but when troubleshooting using the simplest set of settings sometimes helps to narrow things down more quickly.

Set your trigger count to 1 and try to trigger the alert manually by moving your high below the current value.

basically see if you can get it to trigger on a high reading and then slowly work back to your original settings to see where it breaks.

one last thing (this is probably not the case but I thought I'd mention it anyway), I'm not sure how you set up your triggers but depending on the type of graph you may have attached it to the wrong data set (for instance processor load average diaplays 1 min, 5 min and 15 min averages) if you attach your monitoring to the 15 min tab you will not get alerts until the 15 min avg goes above the max which could be a long time into an event.

I feel really stupid but I figured out the problem. I was getting the above threshold e-mails but they were being flagged as SPAM by my e-mail server and put in a different folder. No idea why the above thresholds were considered spam and those below weren't but that was the problem. THOLDS is working perfectly now!!

nice one!

I 'm getting high and low threshold emails. The test e-mail funtion also works. But what does not seem to work is the "Dead Hosts notifications"

And they are not being blocked bij the anti-SPAM software

I have currently added a few extra things that should possibly help in the message not being flagged as spam, someone was kinda enough to post their spamassassin report of one of the emails, and it has helped greatly.