Modernize your API stack with GraphQL and Azure API Management | Azure Friday

[@kevinrobinson676]

Be good to see your implementation Scott, nice simple use case

[@MarvijoSoftware]

Hi Scott. Good episode as always. Why would we move away from OData with Swagger to GraphQL?

[@FelipeLima-vd3um]

Great video guys! Thanks a lot

[@surajbahl-f9m]

In this case frontend has to understand the graphQL schema, how is it different then calling multiple rest api's and aggregate the data? if we are concerned about under or over fetching then it doesn't help if backend is still fetching all data regardless what has been asked in graphQL query.

[@vivekkaushik9508]

Scott I have a question - Why should developers prefer GraphQL over Swagger? Swagger also comes with visual representation of APIs and documentation. Maybe you can invite Elizabeth again to discuss Swagger vs GraphQL in APIM.

[@MicrosoftAzure]

Hi there, Vivek! Swagger (which describes REST APIs) and GraphQL are two different ways of describing APIs. Developers prefer GraphQL APIs over REST APIs for data-driven UI applications because they can describe the query to fetch all the data for rendering a page in a single request. This results in a more efficient transfer of data - only the data that is needed for rendering the page is requested, resulting in less requests and the precise data needed. With REST APIs, rendering a page generally results in multiple API calls and throw-away data - data that is returned in the response but not needed by the page. GraphQL APIs also support real-time notifications, allowing a UI to be responsive as new data can be presented. Having said that, GraphQL is not for every situation; service-to-service communication is generally more efficient over REST or gRPC than GraphQL, and implementation complexity is generally higher with GraphQL than REST, resulting in longer development times and more complex maintenance scenarios. We hope this helps to answer your question!

[@kevinmiles7152]

@@MicrosoftAzure A REST API can also be designed for fetch all the data needed (and only what is needed). A full-stack developer can design the REST API that is needed by the front-end when this is being worked on at the same time. Some Developers might prefer GraphQL , but some Developers prefer REST.

[@vivekkaushik9508]

Just in time when I was looking for some way to make my API more coupled from the client.

[@TheUsamawahabkhan]

What about caching and performance as compared to Rest ?

[@arjain18]

If so much data can be called in one call, will it not become like monolith api

[@vivekkaushik9508]

Protip - Watch at 1.5x

[@Biztalkcontrolcenter]

Why doest he stop talking and let the presenter do her thing. What a chatter box!!