Support for External Cache Invalidation Across Microservices

[AlirezaEiji191379]

Hi, first of all, thanks for this awesome library!

I’m working on a system that has two microservices: A and B.

• Microservice A is the source of truth for a certain entity.
• Microservice B reads that entity from A and caches it using FusionCache.
• Microservice A also uses FusionCache — but it does not cache this particular entity.
• We want B to cache it for performance reasons.

Here’s the challenge:

When data is updated or deleted in microservice A, we want the cached version in microservice B to be invalidated as well.

Since both services are using FusionCache (even though A doesn’t cache this entity), is there a way to trigger a cache invalidation in B via a Backplane mechanism (like Redis or anything else), from A?

I was thinking of using Redis Pub/Sub or a custom webhook from A to B to manually call .Remove() in B — but I wonder if FusionCache’s Backplane could help automatically propagate invalidations across services?

Would love your input on the best way to handle this with FusionCache. Thanks a lot 🙏

[jodydonetti]

Hi @AlirezaEiji191379 , thanks I'm glad you're linking FusionCache!

Regarding your question: are you using FusionCache with only L1 or also with L2?

[AlirezaEiji191379]

Hi @AlirezaEiji191379 , thanks I'm glad you're linking FusionCache!

Regarding your question: are you using FusionCache with only L1 or also with L2?

using L1 and L2

[jodydonetti]

Hi @AlirezaEiji191379

Here’s the challenge:

When data is updated or deleted in microservice A, we want the cached version in microservice B to be invalidated as well.

Since both services are using FusionCache (even though A doesn’t cache this entity), is there a way to trigger a cache invalidation in B via a Backplane mechanism (like Redis or anything else), from A?

Yes! When microservice A update something, instead of not doing anything related to FusionCache, it can do something like this:

cache.Set("Foo", value, options => options.SetSkipMemoryCache());

This will tell FusionCache to update "foo" BUT skip L1 (memory cache) while doing so (since you don't need to cache it there): this in turn will update L2 (distributed cache) and send a backplane notification to all the other nodes.

When microservice B receives the notification, it will do what is described here, meaning:

• if "foo" IS NOT already in the L1 of microservice B, it will do nothing: then at the first request for "foo" on microservice B all the normal flow will happen, getting it from L2 and copying to L1
• if "foo" IS already in the L1 of microservice B, it will immediately update it from L2, so the new value will be available locally right away

Hope this helps, let me know if you try this.