Check process running
This alert means that a process which should be running is not. It’s highly likely that this process corresponds to a service.
If the process doesn’t correspond to a service then it will be necessary to find out more about how the process runs. Read troubleshooting the process at the end of this page.
Check the status of the service
$ sudo service <process> status
If this doesn’t work, you may need to do some digging to find out the name of the service. For example, the
postgresql service runs the
Determining the name of the service
/etc/init for services that may be in charge of the process.
You can also list services with
sudo service --status-all, though this doesn’t show processes started with upstart.
Fixing the issue
If the service isn’t running, it can be enough to just restart the service by using:
sudo service <service> start
If the service is referring to a GOV.UK application, it may be necessary to also restart the Procfile worker:
sudo service <service>-procfile-worker restart
Sometimes, a process might appear to be running, but is actually stalled by a child process that has completed but not been garbage collected:
$ ps -ef | grep defunct root 2735 1 0 Jun08 ? 00:02:09 [prometheus] <defunct>
You can confirm this is a child process of the process in the alert by running
ps faux to see where it is descended from.
sudo service <service> restart should bring the process down, kill of any child processes and start it up again, but if the defunct process is still hanging around (and especially if its parent process is now
1), you may need to reboot the machine.
Troubleshooting the process
If the process doesn’t come back, then it’s more likely that there is something going wrong with it. You can start investigating by looking in the log files which could be in one of the following places:
If you see something in the logs such as
---> Spinning up 'govuk_crawler_worker' (type bare) in 'production' environment /usr/local/bin/govuk_spinup: 30: cd: can't cd to /var/apps/govuk_crawler_worker
This suggests that the crawler hasn’t been deployed to the machine. You can try re-deploying via deploy jenkins.