I disabled the phantomjs cronjob – it’s eating resources and this is a shared server with other services. Each process was small, but it adds up when there’s dozens of them! They were what lead to the server using 100% swap!
Nice idea if Atlas was the only thing running. I have currently removed the cronjob altogether. If you can adjust it so that it only runs a couple times in a given 24h period – I’d say once a day is reasonable. Keep in mind that you have to care about resources of other services. Where are these screenshots even used?
After killing off your cronjob, the server resources are no longer strained. It was bad.
Sharing is caring. The server is shared by the following:
- demo.openmrs.org
- Modulus
- redirect.openmrs.org / om.rs
- errors.openmrs.org
- resources.openmrs.org
- and finally Atlas.