What the page shows

High availability is handled by Proxmox. The HA page reads quorum, manager services, resources, rules and groups live from the selected cluster, so it shows what failover will actually do.

A quorum badge, and the HA services per node. Lost quorum is the first thing to act on.
HA resources with requested state, group and restart and relocate limits; a badge marks resources of the current project.
If HA is not configured, the page says so.

Affinity rules (Proxmox VE 9)

Node-affinity rules pin resources to nodes; resource-affinity rules keep VMs together or apart.
Rules show type, members, nodes or affinity, strictness and comment; disabled rules are flagged.
Add VMs (ha.manage, node-affinity rules only) picks from HA-managed VMs that aren't in the rule yet.
On Proxmox VE 8 the page explains that the rules API doesn't exist instead of showing an empty table.

HA groups (legacy)

HA groups list nodes with priorities, whether resources are restricted to them, and the no-failback flag. They are marked deprecated when affinity rules are in use.

Automatic placement

With HA enabled in a template's Defaults, new VMs are registered as HA resources and added to the node-affinity rule of the node they landed on, detected from the rules' node priorities.
This happens at the end of the deployment so in-guest reboots don't trigger HA, and it never fails a deployment.
Migrating a VM moves it to the target node's rule first. Suspending a VM takes it out of HA until it runs again.

Features on this page

High availability overview
Automatic HA placement