Handling unreported metrics on termination

Hello, I am looking for some guidance on how to handle unreported metrics when a microservice is terminated. Consider a Kubernetes based application that has decided to terminate a pod between scrapes from the server. The terminating service has metric updates that have not been scraped. These updated values need to be reported.

Use of the push gateway is not a solution.

Short of ensuring a clean shutdown as is described in best practices for terminating Kubernetes pods, are there any other solutions that anyone has implemented?