Description
Hi, echoing the thread Azure/durabletask#662 in this repo.
I believe SQL provider works similar to AzureStorageOrchestrationService, which only terminate running orchestration. In our code, we schedule one orchestration which will be triggered in a month. Before it is executed, we also allow user to terminate the orchestration.
However, those scheduled orchestration will be not terminated right away but until the timer expires. Since this is a rather active task for customer, it generates 3K rows in NewEvents table within 3 weeks.
Is this a provider behavior or the core behavior? It doesn't seem necessary to terminate the orchestration only when it starts running. The orchestration instance already presents in the db across the board. If it is provider behavior then I think we can have one improvement of terminate both pending and running orchestration in order to reduce the load in NewEvents table