Hello, We are using the grep_log plugin within an Agent (linux server). We are monitoring a log file (/var/log/messages) for certain messages, specifically we are looking for a line that states "Error connecting to XXX". We have the grep_log plugin looking for those words and it has been working great except for one issue: when the error appears in the log file, we get an alert fired for every occurrence. This quickly becomes a problem because the process on the server that is creating these error messages will repeat the error constantly until the issue is resolved (by restarting the service). It will literally create about 800,000 lines in the log file with the same error over and over, we have been bombarded with emails from Pandora and can't make it stop until it runs its course. I thought by setting the 'threshold' option in the Alert settings to 3600, it would reduce it to 1 e-mail per hour but I believe this will only prevent it from repeating the first email on the first occurrence of the error. Is there a setting I can use that will limit the amount of emails sent? Thanks
Thousands of Alert Emails for Agent Plugin
caesar · 26 nov 2013, 4:42 · 7 mensajes · 65 visitas
- Inicio del debate
Hi The only option would be good is that we only run the first time, but so clear whether this accounts me all the time recovering and returning to critical condition ... Truth?
Sorry, I am not understanding you. What option can I set so that the alert is only triggered on the first instance instead of it repeating itself. Right now, there were 800,000 error messages in the log file, I am still receiving emails from November 22! How can I make it stop?? lol
What I meant is that you set the alarm on the minimum number of alerts to 0, and the maximum number to 1 in the time threshold you want (1hour for example) and thus will receive an alert each time, unless the state to recover in that time, so it will execute before ...
Sorry, I still don't understand your message. Hablo espanol si quieres... Gracias.
When configuring an alert action you have 3 parameters. 1.- Min number of alerts 2.- Max number of alerts 3.- Time threshold To get only send 1 email when the module appears in critical condition for the first time and never again send another until recovers the alert the correct configuration would be. 1.- Min number of alerts - 0 2.- Max number of alerts -1 3.- Time threshold -A very high range, eg one day Thus, if your module is shown continuously in critical condition, because it always sends errors, you will significantly reduce the number of alerts.
Thank you, that seems to have resolved the issue for sending many critical e-mails. This time, it only sent one e-mail for the critical issue however when the issue was resolved it sent out thousands of the "RECOVERED" emails!
Entra en tu cuenta para publicar.
