Inaccurate Vcenter Measurments

[DaemonCypher]

Trying to write a script to detect any VMs with high resource usage. Using the cpu.usage.average and men.usage.average as indicators for finding VM's with high resource usage. However, The numbers I am pulling are not reasonable they are outputting way above the so called 100 percent usage( sometimes 600 -2000). I tried using other metrics to get a more accurate measurement of VM resource usage, but no luck so far.

[scott28tt]

Thread reported to moderators, it needs moving to the vSphere area so it is seen by the correct audience.

[CLINZ]

This may be more than you are looking for but i started on a similar path as you, looking at writing a script and stumbled across a blog written by Jorge De La Cruz and how he used Grafana, Influxdb to monitor vmware. It looked interesting so i thought i'd have a look at implementing the solution. Going slightly off topic but since this little forray in to Grafana we've expanded what we monitor with it.

But... Back to the question. The process would involve setting up a VM and following the tutorial on the blog but could be well worth it if your looking for a dashboard style reporting and monitoring tool.

If not then another option maybe to look at the .json files and have a look at the scripting he has used to pull metrics from vcenter and see if that can help you put the sense back in to the metrics.

Link to his dashboard is below but a quick google can also bring up what you might need.

https://grafana.com/grafana/dashboards/2840

Hope this helps.

[DaemonCypher]

I am just using Pyvmomi community sample vm_perf_example.py code, and that is giving me these odd numbers. I would assume since an open source code would be able to find such bugs early on.