Simple service monitor written in Bash. It just provides the monitoring logic. All alerts, actions and monitor probes are on you.
You can get the syntax on the command line by running ./monitor.bash --help.
Usage: monitor.bash [options] "command_to_monitor"
Options:
--on-init 'command'
Command to execute on monitor start.
--on-up 'command'
Command to execute when service is brought up.
--on-down 'command'
Command to execute when service has gone down.
--rest-time n_seconds
Sets all rest timers to n_seconds.
--rest-time-while-down n_seconds
Sets the rest timer for the down state to n_seconds.
--rest-time-while-up n_seconds
Sets the rest timer for the up state to n_seconds.
--raising-back-rest n_seconds
Sets the rest timer for the raising back state to n_seconds.
--falling-down-rest n_seconds
Sets the rest timer for the falling down state to n_seconds.
--falling-down-attempts n
Sets the number of attempts for the falling down state.
--raising-back-attempts n
Sets the number of attempts for the raising back state.
Command to monitor must return 0 on success or non-zero otherwise.
Monitor will start on the INIT state and assume the service is up. On a failure, it will transit to the FALLING_DOWN state, which will repeat the task N times to make sure this was not a temporary failure. If this is not the case, monitor will go to the DOWN state. On a success it will transit to the RAISING_BACK state and finally, if confirmed, reach back to the UP state.
./monitor.bash --on-down "mailx -s 'DOWN: ping 127.0.0.1' /dev/null" "ping 127.0.0.1"
This program is NOT considered stable. Option parameters may change without notice until proper stabilization is reached. Feedback and patches welcome.
The following help is welcome:
-
A prepared script to send e-mail in a friendly way from the command line.
-
Feedback on option names.
-
SMS scripts for different cellular phone providers.
-
Command line tools to get the state of all running monitors.