I have seen some posts around this subject but none of them did answer our question.
Our Pega Application hosts an API service that is heavily called the moment that a node has started up. This API uses a node level data page containing lots of data, so this takes around 30 seconds to load an a node. As a result the first couple of API calls will time out because of this load.
We want to make sure that a node only becomes available for the load balancer once this data is loaded. How can we do that? So we want to add the node level data page loading as a startup task of the node.
Note there will not be any user login to this node, so activities like ApplicationProfileSetUp will not help us here.
You can solve this by creating a Job Scheduler that runs automatically when the node starts. This scheduler should call an activity that simply references your node-level data page, which forces it to load right away. Once the data is fully loaded, set a flag (like a Dynamic System Setting or a custom data page) to mark the node as ready. Now, set up a custom REST endpoint or healthcheck service that returns a success response only if this flag is set. Configure your load balancer to use this custom healthcheck before sending traffic to the node. This way, the node won’t receive API calls until the data is preloaded and it’s fully ready