Currently, instance types must be defined as "on-demand" or "reservation" at build time and built through different workflows.

If you try to build a reserved instance and no reservations are available, the build just fails.

Would like to see a single workflow where if a server reservation is available, the instance deploys to that reservation.

Perhaps something like a tag in the on-demand workflow such as "use_avail_reservation: true"?

Where false (default) will act as it currently does and true will check for available reservations before building an on-demand instance and false will just build the on-demand instance without checking for reservations?

and/or possibly the reverse scenario for the reservation workflow with a tag like "ondemand_allowed: true"

Where again false (default) will act as it currently does and when true, instead of failing when no reservations are available, the workflow will auto-switch over to use an ondemand instance type. Though this 2nd option is understandably "riskier" from a business perspective as a customer might think they're building a reserved instance and get an on-demand one at a higher bill rate instead.