More specifically, I am talking about Kubernetes pods, running different instances of the same application. I haven't found official recommendations for that scenario.
Besides attaching a Persistent Volume, it feels like I need to also care about the race conditions.
I am sending Application Insights' Events and Metrics to be consumed by an orchestrator service, so I need them to be reliable. Should I expect that if I don't care about inter-processes race conditions, duplicates might appear?