Health monitoring solution from inside of virtual machine.

[NetBear]

Hi all,
I'm taking a part in developing of some application will run on  virtualized platform, currently it going to be on VMWare ESXi but I  would not exclude the KVM or Xen option. The application is real-time and going  to be very sensitive to resource shortage, especially to CPU and  network bandwidth. Because this extreme sensitivity we need some health  monitoring mechanism will be able to gather statistics about the system  resources provided to specific virtual machine with our application  running inside. It may have feature to record collected statistics to  log file or send alarms. It’s not a problem to collect such data from  host (hypervisor) level, but the application running inside VM hasn’t  required permissions, cannot access ESXi API as well.
Any suggestions or ideas would be very much appreciated.

[AndreTheGiant]

Welcome to the community.

Which kind of resource? How do you collect this data?

Andre

[NetBear]

Thanks for welcoming Andre.

Regarding resources, when VM created you set its properties, for ex. 2 vCPU units, 2048 MB of memory, 100 GB of HDD space and finally some network bandwidth. For example, we're supplying our application to some customer and providing exact requirement for VM (see above).

I need some health monitor process installed as part of my application, may be as a part of watchdog module. This process will run inside VM container and perform monitoring of system resurces provided to my application in current moment of time. If customer (IT manager) put 20 VMs on the same server with extreme resource sharing policy and as result at some moment my application failing to provide the service I need some evidence that the cause of the problem is the resource shortage.