Monitoring a VM when Running in Cloud

The basic instance status for the cloud instance information are presented in the Summary page portlet added by the Velostrata plug-in. When the Remote Console field is Ready, the instance is reachable from the Velostrata Virtual Machine on-prem. The remote console probe tests the RDP port (TCP/3389) for Windows VMs and the SSH port (TCP/22) for Linux VMs.

While running in cloud, further performance and operational statistics are available for the VM in the Monitoring > Cloud Instance tab. You can review details about storage IOs in cloud vs. those affecting the vSphere Datastore storage on-prem, as well as other operational statistics such as IO latency and throughput.

Graphs are offered at pre-canned periods of Hour, Day, 7 days, 30 days as well as a user-defined custom range.

The following metrics are available:

  • Workload:
    • IOPS: Storage IO operations that the VM has generated over the Cloud Edge node.
    • IO Latency (ms): Average storage IO operation latency (1 min interval) for this VM’s operations.
    • Throughput (Bytes/sec): Storage throughput utilized for this VM.
  • Uncommitted Writes (Bytes): The amount of write data acknowledged in cloud for this VM, and still pending a commit to the vSphere datastore on-prem.
  • vSphere Datastore:
    • IOPS: Backend IOPs generated against the vSphere datastore for this VM’s activity.
    • vSphere Datastore Bytes (cumulative): Total bytes read from on-prem datastore or written to during this VMs run in cloud. This graph slope can be used to identify active working set converges and transitions.
    • Throughput (Bytes/sec): Total throughput of vSphere datastore storage access, as generated by the VM’s activity in cloud.
    • IO Offload %: Based on the ratio between IO operations handled in cloud vs IO operations generated against the vSphere datastore on prem.
  • Change set Compaction Ratio %: Efficiency metric for overlapping writes compaction. Pre compression.