Caching an existing Graphql endpoint on kubernetes

I have a graphql endpoint on top of postgresql database deployed in K8S.

I wonder if it is interesting in terms of performance to deploy Kong gateway connect it to graphql endpoint and use the caching from it ?

What is your experience on it ? Does it make sense ? What is the effort to deploy it for one endpoint ?
What should I need to look to imagine how much gain I can have ? I have Jaeger tracing in place, I have one span which takes up to 500ms average to be solved.

Thanks for your insights :slight_smile: