Running an On-premise Pipeline from a URL
The On-premises URL refers to the Groundplex machine from which you send the request. Although the URL options are not tied to Snaplex type, these URLs depend on the user's ability to access those networks. For example, the Cloud URL is available to use from any computer with Internet access, whereas the On-premises and Override URLs require network access to the Groundplex or to a load balancer.
By default the Groundplex machine only accepts HTTP requests from the Loopback interface (127.0.0.1). To allow access from other machines, you must change the HTTP interface of the Groundplex configuration to listen on any IP address (0.0.0.0). This may also require additional network changes, such as opening up a firewall, so consult your network administrator for guidance.
When a Triggered Task is invoked through an On-premise or Override URL, the Task is cached for subsequent executions. The cached Tasks are then used on subsequent triggers of the URL, reducing task execution latency. Triggered tasks expire from the cache after an hour.
Additional Setup for Multiple Groundplex Nodes
A Triggered Task can run on any of the Groundplex nodes. Because of internal load-balancing between Snaplex nodes for Triggered Tasks, the node in the multi-node Snaplex designated in the On-premise URLs might not be the node that executes the Pipeline associated with that Task. Therefore, you need to configure the Groundplex nodes to accept requests from the other nodes so that data can be streamed back to the client.
You can also determine if Snaplex nodes are not communicating by clicking Dashboard > Health, selecting the Snaplex, and viewing the Alerts for each node. If there are communication issues, an alert such as Neighbor connectivity alert might appear.
Configure Triggered Tasks for High Availability with Override URLs
Two additional configurable URLs (Local override URL and Local override Secure URL) appear on the Task Details page. The Override URLs differ from the On-premises URLs in that you can use them to trigger a task to run on any Snaplex node, whereas the On-premises URL always runs on the Snaplex node that was specified when the Task was created.
If a Triggered Task is assigned to Snaplex A, then it always runs on that Snaplex if triggered with the On-premise URL. Therefore, even if triggered on a node on Snaplex B, the Task still runs on Snaplex A. However, if the Override URL is used, then the Task runs on Snaplex A if triggered on Snaplex A and Snaplex B if triggered on Snaplex B.
For example, if a Task is configured to run on Snaplex A, then it can be forced to run on Snaplex B by substituting the hostname and port of a JCC node or FeedMaster in Snaplex B, in place of the one configured in the Override URL.
The Override URL was introduced for users interested in setting up their Tasks to run in High Availability configurations. The Override URL triggers a Pipeline to run on any Snaplex, regardless of what is defined in the Task. Unlike the On-premises URL, which always runs on the Snaplex defined in the Task, you are responsible for defining a High Availability architecture, and it is commonly done using a load balancer or the network's DNS.
On the same screen, you can review task status to determine if you need to run a Triggered Task on another Snaplex. Furthermore, clicking on the task detail displays a window with Pipeline statistics, parameters, and information such as path, node, and the user who ran the Pipeline.
