I install process exporter on the server to be monitored and monitor the alive status of the process by comparing the value of the “namedprocess_namegroup_num_procs” metric of the process I want to monitor.
On the server to be monitored, if a process is terminated with a command such as “systemctl stop ” or “kill -p ”, the value of the “namedprocess_namegroup_num_procs” metric is obtained as “0”, and it can be detected that the process has stopped as expected.
However, if the monitored process does not start when the monitored server is started (i.e. the process has not started even once since the server was started), data on the monitored process cannot be obtained. Of course, information can be obtained from process exporter without any problem for processes that started at the same time as the server was started.
Why does this situation occur?
Also, how should I monitor processes that do not start after the server is started?