In the meantime, I made a small library to query the API and made it do HTTP caching for a few endpoints. It makes filtering quite faster because all /build/{id}/gradle-attributes responses will be cached
https://github.com/gabrielfeo/gradle-enterprise-api-kotlin
Gabriel Feo
02/15/2023, 5:10 PM
Haven’t talked about it much but it’s working well internally
m
Mike Cumings
02/15/2023, 5:12 PM
My main problem is filtering the massive result list of build IDs to only the subset that I actually care about without having to make calls additional to additional endpoints. It is these additional calls by the client for the sole purpose of being able to perform filtering which make the API so expensive.
g
Gabriel Feo
02/15/2023, 5:24 PM
I agree it’s a problem. Until it’s addressed, caching isn’t a solution, but can make it less expensive by not having to query the additional endpoints more than once. Another possibility is the internal endpoint that’s mentioned in the thread I linked earlier.
j
Josh Rotenberg
02/15/2023, 5:30 PM
another vote here for aggressive caching. i proof of concepted a graphql implementation in from of some endpoints that knew how to cache stuff so that we can get more focused data and get it faster