Problem: when we grow to having multiple instances to support our burgeoning experiment count, we will run into the issue that only one booking system can manage an experiment. There is no communication between booking services running on different instances, so for a given organisation, the booking service on one instance should be responsible for bookings of all experiments, with the booking services on other instances turned off (or left empty). When it is time to cancel a session that is already started, the booking service will need to send a cancellation request to an off-instance relay service, so an external HTTPS request will have to be made (rather than a local, potentially unsecured one behind the instance firewall).
What's needed: check code support for making HTTPS request, and test that this works in practice
Linked issue [#19] should also be re-opened and fixed as part of this work.
Problem: when we grow to having multiple instances to support our burgeoning experiment count, we will run into the issue that only one booking system can manage an experiment. There is no communication between booking services running on different instances, so for a given organisation, the booking service on one instance should be responsible for bookings of all experiments, with the booking services on other instances turned off (or left empty). When it is time to cancel a session that is already started, the booking service will need to send a cancellation request to an off-instance relay service, so an external HTTPS request will have to be made (rather than a local, potentially unsecured one behind the instance firewall).
What's needed: check code support for making HTTPS request, and test that this works in practice
Linked issue [#19] should also be re-opened and fixed as part of this work.