Host Monitoring- CPU Metrics?
I have couple of questions regarding the metrics reported by the new host monitoring agent. I have looked at the documentation in detail but was unable to find an answer:
Are the Host CPU Utilization metrics reported by dynaTrace based on the virtualization aware timers? Meaning if I have multiple logic partitions (LPARs) and they are all sharing the same CPU resources and a fixed entitlement policy is defined for each one of them, how would dynaTrace host CPU Utilization metrics takes this into account? Will it be able to highlight for e.g that a specific LPAR is going beyond it's CPU entitlement?
Any thoughts or feedback would be appreciated.
Thanks,
Jigar
Host Monitoring Agents are currently not supported in dynaTrace 4.2. Check out the full overview of supported metrics per OS in the online doc: Host Health Monitoring
Thanks Andi. To be more specific:
My question was regarding the host metrics that are pulled automatically when we have the Java agents installed (introduced in 4.2). This is really very helpful as we get these host metrics without any extra configuration. The questions I raised was regarding the Host CPU Utilization metrics reported in the virtualized environment as explained above.
Hope it is more clear now.
Jigar
Our Agents (Java, .NET, Web Server, Native or Host Monitoring Agent) use Native OS API calls to obtain that information. The measures reported should therefore be the same metrics as the OS specific monitoring tools provide for that particular OS Instance.
I hope this makes sense
Thanks Andi. I am interested in knowing which native OS API calls are made for AIX OS?
Jigar
We use perfstat to capture this information
