I am developing a REST client to interface with RHQ 4.4 to create a custom display of 100+ agents but I am having some difficulty getting the necessary data from the REST interface.
I am trying to view alerts fired by child resources (file system, network adapter, DB) for a platform through the REST interface but I am only seeing alerts for the platform's schedules (CPU, memory, etc...). Are alerts from child resources supposed to show in the REST alerts list?
Also is there a way to customize the interval used to calculate the average for each metric? I have a requirement in my project to to use the average of the last 15 minutes to smooth out sudden fluctuations in the data, which I am currently doing by just using the raw metrics but if the average on the metric response had this custom average it would cut out a lot of processing in my app.
In addition to customizing the metric averages I would also like to set alerts on the custom averages. I have tried setting alerts on threshold values by setting the dampening to account for the number of metrics that should be returned in 15 minutes but the number of metrics returned does not always match the schedule. All of my metrics are scheduled for 30-60 second intervals but sometimes the metrics scheduled for 30 seconds only show up every 60 seconds so alerts that should have been triggered are not because the alert definition is expecting 30 metrics in 15 minutes but only get 15. Is there a way the ensure the agent is reporting every metric as scheduled or setting up an alert for a custom average described above?
As a note I had read somewhere that an agent will delay metrics if it is unable to gather/send them all at a collection interval but I am seeing these delays on my development machine that is running both the server and only one agent and the CPU and memory utilization are always <5% and <20% respectively.