respond stale cache if L1 cache has expired and origin backend is down

[estebanlartigue]

Hi, is there an out of the box stale cache for unhealthy system?

something that we can use whenever our L1 cache is expired and then the origin backend system is down ?

I saw in another question something like using a duplication of the response cache but without expiration time. is that a good thing? I thought L1 cache in apigee TTL are not super enforced since they are going to be removed when..... apigee needs the resource.

any thoughts here?

my goal would be

1. find out if apigee has the concept of "stale" cache use if error from backend is XYZ.

2. if not, what is the best way to produce this stale cache. is this having dups of cache with different TTL ?

[cjking]

> 1. find out if apigee has the concept of "stale" cache use if error from backend is XYZ.

No such concept exists by default.

> 2. if not, what is the best way to produce this stale cache. is this having dups of cache with different TTL ?

Apigee doesn't recommend caches > 30 days: https://docs.apigee.com/api-platform/reference/limits

The KVM is an option, if the amount of data you're looking to store has few enough entries and each entry is small enough.

The other option worth considering would be using an extension to interact with a datastore, such as https://docs.apigee.com/api-platform/reference/extensions/google-cloud-firestore/google-cloud-firest... but you should keep in mind the recommended max 120 TPS on extensions (see the limits page)