Is mesh response-cache plugin compatible with query batching?

[hcentelles]

I’m asking here instead of opening an issue because it’s not clear to me whether all these components are supposed to work well together. I'm using Cloudflare KV cache storage.

• I have a GraphQL source with a schema containing query A. I extend that schema with additionalTypedef and additionalResolvers to add query B.

• The response cache plugin is set with ttl=0 except for query A, which uses ttlCordinates.

• I sent queries A and B in a batch request. The response returned both results correctly, missed the cache in that request, but cached the response.

• Then I sent the same A and B queries in a batch again. This time, the response was the results of query A repeated, instead of the results of A and B. The cache hit was true. Like this: