High Availability

HA configuration ensures your services stay available.

In standard configuration, if your server or Agent is unavailable, the integration service will be blocked. In order to avoid down time for your internal and external customers, High Availability configuration can be set up to ensure there will be Agents available to process the requests.

Added bonus of High Availability is the capability to scale horizontally, allowing for more Process executions within the same time window.

High Availability Configuration

In HA configuration, multiple Agents are placed into the same Agent Group, so that they can distribute the Process executions between themselves.

In order to coordinate this workload sharing, a load balancer and a shared database are required for HA configuration. Without the coordination, essentially only one of the Agents within the Agent Group will be executing the Processes, known as Primary Agent in that scenario. Extra setup is thus required before HA configuration can be achieved.

Depiction of High Availability configuration.

Once the configuration is set, the workings of multiple Agents performing the same tasks will become mostly invisible to the end users and developers. In general, developers do not have to worry about the Agent or Agents as they are, instead they are mostly working with Environments and Agent Groups, which already abstracts the Agent performing the actual executions. Processes and APIs are always deployed to an Agent Group and not to a specific Agent, so nothing really changes for the users in Frends.

To see that HA configuration is operational and sharing the workload, you can see this by viewing the Process Instance list for some Process in the HA Agent Group. On the list it will detail, which Agent triggered the Process for each execution, making it visible that different Agents are running the same Process side by side.

Apart from the triggering Agent in the Process Instance list, HA configuration is invisible to users.