hat "rare" bug that looks intermittent but is actually deterministic. It is just waiting for your pods to scale.
I recently dealt with a textbook concurrency issue in a Java application running on multiple parallel instances, and the solution turned out to be much simpler than the workarounds we had piled on top of it. This is a walkthrough of the problem, the failed attempts, and the final fix, with code and diagrams so you can apply the same pattern.
The Setup
The application stack:
. Java with Quarkus and Hibernate (Panache)
