I am trying to determine how triggers behave. I have a trigger set up for Server Unreachable. I have the delay in between messages set to 30 seconds. If I run zabbix and then kill the client machine, I get an alert message as expected. Then I get two more thirty seconds apart. Then I get no more. What causes it to stop sending messages? I would expect to either get only one message when the client machine goes down, or keep on getting messages as long as the client is down. Does anyone have an answer?
-
-
Interesting. I've never been clear what the delay between messages value does. I'm using Alpha10, and I only get one message when a trigger goes off. What version of Zabbix are you using and what do your triggers look like?
Nate -
I am using alpha 10 also. My trigger is this: {linbox:status.last(0)}=2
My action for this trigger looks like this:
Send message to: single user
User : Admin
When trigger becomes: ON
Delay between messages (in sec): 30
Subject : Server {HOSTNAME} is unreachable
Message : "147.24.71.12
status
{linbox:status.last(0)}"
Scope : This trigger onlyComment
-
Because Zabbix has only ever sent one message per alert, I've always thought that the delay between messages refered to the time between sending a message to a user with one type of media, and sending the same message with the next type. Or possibly the time between messaging on user, and the next user.
What media does your user Admin have associated with them? If they have several media types, then they would get a message for every media defined, and my theory about delay between messages would make sense there. If they don't have multiple media types, then I don't know why you're getting several messages.
What happens if you set the delay between messages higher? Do the several messages get spaced out farther? Your trigger settings seem normal to me. Does this happen for all your triggers? Does this happen for other users you have defined (if any)? Just trying to troubleshoot,
NateComment
-
I figured out what was going on. It was my problem. I have some custom code in the email alert section that writes a message to a posix port instead of sending an email. I was recieving three messages on this port 30 seconds apart.
What was happening was, since the email was never being sent, it is hardcoded to retry the email two more times. So this is where the extra messages were coming from. I just had to tell it that the email send was sucessful no matter what. Now it is behaving correctly (only 1 message is sent).
Thanks for the help.Comment
-
I assumed that the time between emails was to eliminate being spamed with alerts if a trigger is flapping constantly.
Admittedly I've never tested this though.Comment
-
Ah, that would make sense. I've never run into that situation (that I can remember), but that would be a smart solution if I did.
NateComment
Comment