|Subject:||Re: Monit not detecting service failure, reports service is up when it is not.|
|Date:||Sun, 17 Apr 2011 12:06:26 +0200|
There are few possible problems:
1.) monit tests in cycles, whereas the interval between cycles is given by the "set daemon <seconds>" in monitrc => in the case that your interval is large (lets say 5 minutes), then monit won't detect the problem until next cycle and may present the process as running in the GUI (this information is cached until next test cycle). Using short interval such as 5s is better to get quick reaction on problems.
2.) The pidfile based test does check for the process running with given PID - you can use the match based check to make sure that the specified process is running independent of pidfile (if you want to use this check, use Monit 5.2.5 or newer):
check process apache matching "/usr/sbin/httpd"
3.) The process uptime was displayed based on the timestamp of the pidfile for Monit < 5.2. The Monit 5.2 or newer shows real uptime (based on process table informations).
=> i suggest to upgrade Monit to 5.2.5 and use match based process check if the pidfile based check is not reliable in your environment
On Apr 16, 2011, at 7:43 PM, Eduardo Gutierrez wrote:
|[Prev in Thread]||Current Thread||[Next in Thread]|