High Availability and Scale Deployment considerations

For Production environments, ensure that all components and its dependencies are deployed in a High Availability architecture. These includes:

• API Instances:

  • Deploy with a Load Balancing front end

  • Have at least 2 machine instances running the API services

  • You can configure Load Balancing with several methods

    • IIS Web Server with ARR as mentioned here

    • Use Windows Server Network Load Balancing

    • Or use dedicated load balancers like F5.

  • Though the API calls are stateless, in order to give a better experience and performance, configure load balancing with sticky sessions (requests from the same client/IP during the same session goes to the same backend API node).

• Agent Services (Windows NT Service)

  • Install Agent Service on the first machine, configure and ensure its running. The first machine will become the primary (master).

  • Install Agent Service on a second machine, configure and ensure its running. The second machine will register itself as a secondary and will take over as primary (master) if the first agent stopped.

• SQL Server

  • Deploy with Always ON / Fail over cluster.

  • Read more on SQL Server Business continuity overview here.