Hi,
In the Cluster - Monitor - Performance section, what is the difference between vSAN - Virtual Machine Consumption and vSAN - Backend, aren't they referring to the same vSAN storage?
How does this help with monitoring and in case there is a latency or throughput issue which one should be the one to start with?
Thanks
Dear andvm You have a very valid question , as it can be tricky to understand difference between Virtual Machine Consumption and vSAN-Backend.
To simplify, please find difference between these two in laymen terms.
1) Virtual Machine Consumption : Every VM which is hosted on a vSAN DS is stored as an object, and Virtual Machine Consumption graph shows the performance of the virtual hardware connected to any VM hosted on the vSAN DS.
2) vSAN-Backend - As we know the vSAN is a SAN created using multiple HDD (SSD or Hybrid) connected to ESXi hosts, to monitor performance of actual HDD this graph is used.
Both sets of graphs cover the statistics for:
So in Simple terms Virtual Machine Consumption graph will show virtual hardware performance and vSAN-Backend will show the performance of actual hardware which is underneath vSAN DS.
Now depending on your use case you can decide which one is worth checking, you will always start with the VM consumption graph and move to Backend after tracing where those objects are stored by vSAN algorithm.
Regards
Pradhuman
VCIX-NV, VCAP-NV, vExpert, VCP2X-DCVNV
If my Answer resolved your query don't forget to mark it as "Correct Answer".
Hello andvm
Essentially the distinction here is between the usage/performance from the Guest-OS (VM) perspective and the vSAN-cluster's perspective.
Further information on these are documented here:
https://kb.vmware.com/s/article/2144493
For troubleshooting we would ideally initially look at both levels - an issue at the cluster or host level (e.g. networking, controller, Disk-Group, disks) can of course cause an issue at the VM 'client' level, but there can be issues that are not caused from a lower-level and are purely at the VM level (e.g. latency under specific IO profiles, using inadequate SCSI controller or networking configuration, problematic Guest-OS).
Bob