Dariusz Kuc

🤔 Curious what it really takes to scale a graph across brands and teams? Start with strong platform visionaries—and a community eager to learn from them. This week's Apollo Community Meetup in Chicago delivered just that. #ChicagoTech #ExpediaGroup #GraphQL #APIOrchestration

Join us for the @apollographql and @ExpediaGroup #GraphQL in-person meetup in #Chicago this Wednesday! See you there!

We're headed to Chicago on May 21—and you're invited! Join us for the Apollo Community Chicago Meetup and get an exclusive look at how Expedia Group scaled graph-based API orchestration across their global brands: apollographql.com/events/apollo-… #ChicagoTech #DevCommunity #TechEvents

Deep dive into @apollographql's Connectors evolution by Lenny - it's a fascinating story of bringing declarative API orchestration to everyone. From a prototype gRPC connector to a full-fledged solution built right into Federation. Love seeing the product journey!

#GOTOchgo '24 was a blast! Check out my #GraphQLFederation talk to see how it can help you build flexible and resilient API strategy that evolves with your business needs. #GraphQL #apiarchitecture

@apollographql router v2.0 is out! Comes with native QP and connectors GA! The best part? You can try it/test it all out for free. Check out the blog post for details -> apollographql.com/blog/whats-new… #GraphQL #GraphQLFederation

@__xuorig__ @TheWorstFounder @apollographql During minification we basically just iterate over the selection set twice -> once to capture how many times given subselection occurs in the query and then we iterate again to lift up subselections that were used more than once.

@__xuorig__ @TheWorstFounder @apollographql Indeed it is a tradeoff between simpler query planning algorithm and the overhead of extra minification. While overall CPU usage will vary greatly based on the underlying query, in general QP algorithm is much more complex than the minification process.

@apollographql router v1.60.1 was just released 🚀! It comes with an improved algo for minifying the #GraphQLFederation subgraph fetch operations. Improved fragment auto-gen now extracts named fragments from all common selections within #GraphQL operation (vs just inline frags).

@TheWorstFounder @apollographql disclaimer: I haven't looked much into it (and whats the root case) but in few cases (where gen results in bigger queries) the diff in size seemed to be caused by the diff between original and normalized query shape (guessing it was from abstract vs concrete type selections).

@derek_kuc @apollographql What is smarter query normalization?

@TheWorstFounder @apollographql *this also yielded a significant performance improvement over trying to re-use user provided fragments

@TheWorstFounder @apollographql Our original approach of extracting inline fragments (released over a year ago ;)) was very simple and naive (and fast!) approach but it did yield better results in 77+% of the test cases (and was significantly decreasing most of the large operations).

@TheWorstFounder @apollographql New v2 gen fragment algorithm is indeed similar to what you've done so the end results should be very similar (it improves query size in 95+% of cases vs hand crafted query). I think further improvements can be gained from smarter query normalization.

@derek_kuc @apollographql Cool stuff! Looks very similar to what we published half a year ago. 😅 Have you found any improvements to what we're doing or is the compression ratio similar? wundergraph.com/blog/graphql_q…

@apollographql Check out router v1.60.1 today! github.com/apollographql/…

@apollographql Auto fragment generation can be applied on all operations regardless whether your original operation used fragments or not. While the end result is dependent on your operations, in our testing over 2.8+ test operations in extreme cases we saw 99+% size reduction! (4.2MB->6KB!)

GraphQLConf 2025 CFP is now open. Submit your talk ideas by May 4!

Curious about #GraphQL Federation? Check out this great intro to #GraphQLFederation by @kamilkisiela -> graphql.org/learn/federati…

Still using #GraphQL #ApolloFederation v1 supergraph? Federation v2 brings tons of improvements and upgrading should be easy. Make sure to update to Federation v2 ASAP! See blog post for details: apollographql.com/blog/migrate-a…

@apollographql native query planner is now officially GA! It's been battle tested with over 700 million ops. The results? 10x (median) faster QP, 2.9x less router's CPU utilization and 2.2x less router's memory usage! Try out the v1.59 router today! #GraphQL #Rust #Federation

Have you tried the new Rust Query Planner yet? Newly released @apollographql Router v1.58 is our LAST release before we go GA! Try it out today -> apollographql.com/docs/graphos/r… #GraphQL #GraphQLFederation