Skip to content

Proactive HA Description

bmwl edited this page Jul 13, 2026 · 1 revision

VMware Proactive HA Overview

What is Proactive HA?

Proactive HA is a VMware vSphere availability feature that prevents virtual machine (VM) downtime by anticipating server hardware failures. Instead of waiting for a physical host to crash and rebooting VMs on other servers, Proactive HA detects early warning signs of hardware degradation and migrates workloads away before an outage occurs.

How It Works

  1. Health Monitoring: Vendor software constantly monitors hardware components like fans, power supplies, and memory.
  2. Degradation Alert: If a component begins to fail, a specialized software plugin (the Hardware Health Provider, us in this case) flags the host's health status as degraded or partially degraded to vCenter.
  3. Automated Migration: Based on pre-configured policies, vSphere automatically uses vMotion to live-migrate VMs off the failing host to healthy servers in the cluster, completely eliminating user downtime.

Requirements

  • vSphere Licensing: Requires compatible vSphere licensing.
  • vSphere DRS: Distributed Resource Scheduler (DRS) must be enabled on the cluster to automate VM vMotion migrations.
  • Hardware Provider Plugin: A provider plugin installed in vCenter to pass hardware health states to vSphere. That's us!

Why Write a Custom Provider Module?

Official hardware health providers are only developed and maintained by major, officially certified vendors (such as Dell, HPE, and Cisco), and are usually only able to be used alongside expensive or complex software suites.

So we decided to write a custom provider module since our infrastructure monitoring used Icinga, and it seemed like the natural place to coordinate Proactive HA state changes from.

By writing a custom provider using the vSphere Web Services API, we are able to manually inject health states (Green, Yellow, Red) into vCenter based on arbitrary custom telemetry, allowing any hardware platform to trigger Proactive HA migrations.

States are managed via Business Process entities, to allow arbitrarily complex combinations of health states (eg. redundant AC, UPS, motherboard sensors, water intrusion sensors, etc can all be rolled up however is needed for that specific environmant)

Clone this wiki locally