cloud native env question

[guykatz]

Hi all;
when using cloud native env, a micro service can be spawned multiple times, depending on threashold defined.
so a specific micro service in a deployment can have 5 instances (5 pods->1 container per pod).
when persisting to microstream in an on premise cloud, for example, I usually persist the data to a volume that all services of the same type are mapped to. the services are not aware of each other and so they all try to persist and read from the same volume, essentielly loading the root object from the same place and trying to manipulate it in memory and store it which leads to multiple writes...
I cant seem to undestand how to bypass this limitation in micro stream (no multi writes).
did I not undertsand something? maybe I missed something?
thanks!
UPDATE: seems to be closely related to this discussion - still, in cloud native this is a basic usecase... and microstream has integration with new frameworks like helidon, micronaut and quarkus does not make sense to me as ReplicaSet is a very common k8s unit to use