Are alerts supposed to run by themselves?

[eloyd]

Just wondering if the alerts are supposed to run by themselves at a regular interval or if we have to kick 'em off somehow (cron, etc). We added one alert that checks every five minutes for a threshold value, but it only seems to run if we run it manually. Works great if it runs.

Would love an option that says "don't notify if the thing is OK." We're not sending it to Nagios yet, just using email notifications, and it seems to send an email in the OK state every time we run it (manually).

[slansing]

Have you made sure that the check is active? There should be a little bell icon to the right hand side of your alert check in log server, that should not have a slash through it.

[lmiltchev]

Would love an option that says "don't notify if the thing is OK."

Actually, there is a check-box on the bottom - "Only alert when Warning or Critical threshold is met.", which *should* do the job, but it's not working as expected as of now (at least when you send an alert manually). Our developers are digging into this as we speak.

[eloyd]

We deleted everything and moved to a new box. I don't recall there being a slash through the active box (and I know I put one there once and then removed it). Will let you know how things go once we get new box all settled.

Didn't see the box that you mentioned regarding run/don't run, but will check later. Time to get pizza before traffic starts up...

[sreinhardt]

Sounds good, let us know how it goes! I had a bit of difficulty at first too, wondering why OK states were not coming in. As slansing mentioned, there should be a "send only when threshold is met" option. I believe this is selected by default and was definitely what I needed to turn off.

[eloyd]

I just went back and and took a good look at this. "Only alert when Warning or Critical threshold is met" is only available if you're editing alerts, not creating new ones. I believe it should be on the new alert input screen as well and would consider this a bug.

[tmcdonald]

Must have been fixed in a revision because I have that option upon creation. It only appears though once you select a dropdown option.

[eloyd]

Ah yah, I see it now. Didn't catch the scroll bar on the window. Okay, thanks.

[scottwilkerson]

[eloyd]

Okay, now I'm just getting angry.

I've got the following set up as an alert (set to 1m intervals just for testing).

alert.png

Notice that it is set to only alert when warning or critical is set to 1 or 2. Notice that it says it's in an OK state.

alert2.png

Yet this thing is still notifying every minute (and so is the other one I made for testing) that things are OK:

Code: Select all

SIP Probes came back with a OK state at Tue, 11 Nov 2014 17:52:32 -0500

The alert was processed with the following thresholds:
- Lookback Period: 5m
- Warning: 1
- Critical: 2

Here is the full alert output:
OK: 0 matching entries found |logs=0;1;2

See the last 5m in the dashboard:
http://127.0.0.1/var/www/html/nagioslogserver/www/index.php/alerts/show/AHglGSbBQW6wzjrYh2SpMg

Nagios Log Server

Please make it do what I think it is supposed to do, which is only to email if it's not OK!