How Does HybridCache Handle In-Memory Cache Invalidation Across Multiple Microservice Pods?

Hybrid caching in .NET 9 is designed to give you the best of both worlds—the raw speed of an in‑memory cache and the consistency of a distributed cache—but it does not “magically” synchronize the in‑memory caches across multiple pods. Each pod still maintains its own local in‑memory cache, and if data is changed in one pod, that update only immediately affects that instance’s in‑memory cache.

The intended pattern is that the distributed cache (for example, Redis) acts as the source of truth. When a change occurs, the distributed cache is updated, and subsequent requests in any pod can then retrieve the refreshed data. Meanwhile, the local in‑memory caches are typically populated from the distributed store on-demand. If you need to invalidate in‑memory caches proactively across pods, this isn’t handled automatically by the HybridCache library. Instead, you’d need to implement an extra layer (such as a pub/sub or messaging mechanism) that broadcasts cache invalidation messages to all pods so they can clear or refresh their in‑memory entries.

In summary, the HybridCache library:

• Provides fast local in‑memory access while still using the distributed cache for global consistency.
• Leaves cross‑pod in‑memory cache invalidation as a responsibility for the developer—either through a messaging backplane or by ensuring that the distributed cache value is always checked as needed—so that stale data is minimized.

This design lets you benefit from in‑memory performance while relying on your distributed cache’s features for data consistency in multi‑pod scenarios.