Citoid config file has been updated via puppet but the code hasn't been automatically updated causing the service to go down (T95616).
For Parsoid and Cxserver, we have set the beta cluster instances to be Jenkins slave, we have a job running after a merge in master happened which checkout the change and reload the service.
The jenkins-deploy user has a sudo policy on wikitech to be able to reload the service.
We should probably set the citoid machine to be a Jenkins slave and craft a job for it.
With the number of services running, we might want to later on migrate to Trebuchet.