Problem

Administrators want to know how long users spend waiting for workspaces to startup. Between the user clicking "start" to the agent accepting connections with startup scripts completed. Measuring this allows template administrators to determine the user experience of "requesting" a workspace and understanding the effects of template changes, provisioner queues, and accessibility of compute.

A single test of workspace startup time does not account for potential intermittent inhibitors to startup time. Administrators would feel more comfortable with a continuous measurement of startup time across the organization.

Proposal

A unified prometheus metric that represents the complete workspace startup time including provisioning and startup scripts. These are available as independent metrics but a customer has directly requested unification.