7. Availability

This page is still in draft.

We covered in Part 1 Chapter 1 that Availability can impact Performance and Capacity, but need to be measured separately.

Horizon consists of different components, each having their own availability metrics.

Object	Availability Metric
RDS Farm	Number of RDS Hosts in Bad State Aim for this number to be 0 at all times.
VDI Pool	Number of VMs in the pool that is not used yet not available. The metric Desktop in Bad State can be unavailable because they are in one of these states: Provisioning \| Customizing \| Starting Up \| Deleting \| Waiting for Agent Maintenance mode \| In Progress Agent disabled \| Agent unreachable \| Agent needs reboot Invalid IP \| Protocol failure \| Domain failure Configuration error \| Provisioning error \| Error Unknown For the description, see this.
Connection Server
Pod	Number of Connection Servers that are not available. The formula depends on the design. If it is N+1, then the availability will be affected when there is >1 unavailable server.

This page was last updated on July 25, 2021 by Stellios Williams with commit message: "Updating URL to Pillar Process People"