Alerting "Baskets"

We are testing New Relic to monitor several remote Linux servers. The application the servers host rely on a myriad of Linux processes running. When in development, some of these processes have been known to crash; the application is then partly functional, and in some cases may appear to be fully functional to the end user (but isn’t really).

We’d like to create a kind of alerting “basket” of predefined Linux processes, as we now do in Nagios. When the process fails, or the number of identical processes (think httpd) exceeds certain high-low range, we’d like to get an alert.

Is this something we can do with the new Alert beta?


Hi @mark_stone

You can alert on the number of running instances of a process by alerting on a specific metric. In this case the metric is formatted as ProcessSamples/ServerName/ProcessName with function instance_count.

Although we do not have a way to monitor a specific PID you should be able to utilize the instance count to craft an alerting condition in our alerts beta that will suit your needs.

This forum post should help provide you with more information about alerting on a specific server metric.