Menu
▾
▴
#1538 [HHQ-2081] Add additional metrics for ESX 3.x
http://jira.hyperic.com/browse/HHQ-2081
VMware ESX 3.x Metrics
ESX Guest VM Service (3.x)
VMCPUReady - This is an important measure of how the VM is performing, This nnumber represents the amount of time that the VM was ready to run on a cpu, but it had to wait.
VMConsumedMemory - This is a more realistic number exposing how much physical memory the VM is actually using. Basically an aggregate number of active,shared, and overhead.
ESX Server (3.x)
HostDiskUsage - This would be an aggregate number for each Target (LUN), and allow us to see if and when we were getting overwhelmed with IO on any given LUN or HBA.
HostNetworkUsage - Similar to Disk only for Nics
HostMemoryBaloon, and Swap - Currently Hyperic gets memory used by VM's if the total were greater than the amount of physical ram it would be nice to know how much is baloon and how much is swap.
HostActiveMemory - This is the number that I use to determine How close to capacity we are on Memory for the host. The amount of memory given to the VM's may be very high or overcommitted but the actual active working set fairly small so this lets us know if we can overcommit more or not.
HostSharedMemory - How much of the Memory Used by VM's is shared. Another metric that helps determine if we can overcommit more. This number also lets me track how long it takes to reclaim memory that can be shared and what will happen in the event of a system reboot that would require all of the vms to reboot. Such as during monthly patching. During the month we may have about 9-12 GB of memory shared and that number will drop to about 3-5 GB for several hours or possibly an entire day after patching and all vms get rebooted.
ESX VM NIC Service (3.x)
Totals for the nic - Right now to get a throughput number you have to look at transmit and receive then add them together. would be much nicer if this were available
ESX VM Disk Service (3.x)
Totals for the Disk - This is the same as for the NIC.
Discussion
Anonymous
