I have a middleware application that sends requests to other services (could be WCF or REST) and the request completion time is variable.
If in case the application is performing a call that takes long time to complete and a SIGTERM is received, is it possible for the app to halt the pod termination until the operation can be completed?
I understand that Kubernetes has a terminationGracePeriodSeconds setting but what if the call to another service takes longer time than that and I don't want the app to be terminated immediately?
I also read that IApplicationLifetime.ApplicationStopping can be used to delay shutdown but I believe eventually a SIGKILL will be issued once a certain time is reached?
Thanks.