This is more about platform memory alerts. From my understanding, we have alert metrics in place that trigger an email if any of the nodes hit the specified threshold in the manager.
However, I am looking at a specific use case. Consider an Ultra Pipeline that needs to invoke a child pipeline for transformation logic. This child pipeline is expected to run on the same node as the parent pipeline to reduce additional processing time, as it is exposed to the client side. Now, if the child pipeline fails to prepare due to insufficient resources on the node, no alert will be generated since the child pipeline did not return anything in the error view.
Is there any feature or discussion underway to provide priority notifications to the organization admin for such failures?
Task-level notifications won't be helpful as they rely on the configured error limits at the task level. While I used the Ultra Pipeline as an example, this scenario applies to scheduled and API-triggered pipelines as well. Your insights would be appreciated.
