Scala:Applied Machine Learning

Process monitoring

The tools described in the previous section are not application-specific. For a long-running process, it often necessary to provide information about the internal state to either a monitoring a graphing solution such as Ganglia or Graphite, or just display it in a servlet. Most of these solutions are read-only, but in some cases, the commands give the control to the users to modify the state, such as log levels, or to trigger garbage collection.

Monitoring, in general is supposed to do the following:

• Provide high-level information about program execution and application-specific metrics
• Potentially, perform health-checks for critical components
• Might incorporate alerting and thresholding on some critical metrics

I have also seen ...