Share
Embed
Rapid Development Techniques for APIs with GraphQL on AWS [Graph Meetup @ Milan]
Alex Casalboni
May 07, 2019
Transcript
54 slides
-
Alex Casalboni Technical Evangelist, AWS @alex_casalboni Rapid Development Techniques for
APIs with GraphQL on AWS @ 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved bit.ly/graphql-aws-milano -
About me • Software Engineer & Web Developer • Worked
in a startup for 4.5 years • ServerlessDays Organizer • AWS Customer since 2013 -
Today’s mobile experience is lacking Mobile website, no native apps
Slow to book, low conversion rates Backend built for high-speed desktops ! -
Your job is to define tomorrow’s experience Digital agency hired
for design First prototype in 4 weeks Must use real data, no mockup ! -
Our Goal Enable backend services for prototype Allow design to
evolve Keeping up with day-to-day job ! -
{ "_id": 3, "name": "ElastiLodge South Zanetown", "image": "https://placehold.it/300x300.png", "description":
” ... ", "location": "South Zanetown", "manager": { "name": ”Marco Rossi", "email": ”[email protected]" }, "timeZone": "US/Central", "phoneNumber": "802-918-3262", "category": 5, "address": { "street": "410 Simonis Forest", "zip": "07545-2308", "city": "South Zanetown", "country": "Virgin Islands, British" }, "deskHours": "24 hours", "amenities": [ ... ], "restaurants": [ ... ] } Legacy REST API -
/posts /comments /authors REST API Chatty & Over fetching {
"posts" : [ { "id" : 1, "content" : "the quick brown fox jumps over the…", "author" : "/authors/123", "created" : 1550486509, "comments" : "/posts/1/comments", ... }, { "id" : 2, "content" : "the quick brown fox jumps over…", "author" : "/authors/123" "created" : 1550486509, "comments" : "/posts/2/comments", ... } ] } -
GraphQL is a query language for your API, and a
server-side runtime for executing queries by using a type system you define for your data. https://graphql.org -
Why GraphQL? posts comments authors GraphQL API Type safe Contract-driven
Efficient Easy to test Realtime -
How does GraphQL work? type Query { getTodos: [Todo] }
type Todo { id: ID! name: String description: String priority: Int duedate: String } Model data with application schema query { getTodos { id name priority } } Client requests what it needs { "id": "1", "name": "Get Milk", "priority": "1" }, { "id": "2", "name": "Go to gym", "priority": "5" },… Only that data is returned -
GraphQL Operations Queries Read data Mutations Write data Subscriptions Listen
for data query { search(q: “harry potter”) { title } } mutation { create(title: “new book”) { id } } subscription { onCreate { id title } } -
Is REST dead then? When data drives UI Structured data
Query-driven Client-driven development Use GraphQL When you leverage HTTP Caching, Content types Hypermedia For resources (e.g. Amazon S3) Use REST Pick the appropriate protocol for your use case -
Welcome AWS AppSync! Managed & serverless Multiple data sources Data-sync,
real-time, and offline Conflict detection in the cloud Enterprise security features -
Data Sources REST Endpoint /graphql Amazon Cognito Amazon DynamoDB AWS
Lambda Amazon Elasticsearch Amazon RDS AWS Lambda -
Resolvers Amazon DynamoDB /graphql Request Resolver Maps payload to underlying
data source via Velocity Template to execute query / mutation. Response Resolver Maps response from data source to GraphQL response via Velocity Template. -
Hotel Detail: Design What data is required? Can we use
existing data sources? How can we enable via AppSync? Challenges? Missing data? -
Mapping to Legacy API { "_id": 3, "image": "https://placehold.it/300x300.png", "name":
"EL Resort & Spa", "description": "...", "location": "Fiji", "manager": { "name": "Florentino Jacobi", "email": "[email protected]" }, "timeZone": "US/Central", "address": { "street": "123 Main St", ”postalcode": "07545", "city": "Momi Bay", "country": "Fiji" }, "phoneNumber": "1-888-123-4567", "category": 5, "deskHours": "24 hours", "amenities": [ ... ], "restaurants": [ ... ] } -
Hotel Detail: GraphQL Schema type Hotel { hotelId: ID! name:
String! image: String location: String! phoneNumber: AWSPhone! address: Address! ... } type Query { listHotels(offset: String, limit: Int) getHotel(hotelId: ID!) } -
Hotel Detail: Resolvers ## Resolver: Request Mapping ## { "version":
"2018-05-29", "method": "GET", "resourcePath": "/hotels/${ctx.args.hotelId}", "params": { "headers": { "Content-Type": "application/json" } }, } ## Resolver: Response Mapping ## #if($ctx.result.statusCode == 200) ## If response is 200, return the body. ## Could also filter result... $ctx.result.body #else ## If not 200, append response to error $utils.appendError($ctx.result.body) #end -
Hotel Detail: Result query { getHotel(hotelId: 3) { name location
image phoneNumber address { street city } amenities } } { "name": "ElastiLodge North Rodgerstad", "location": "North Rodgerstad", "image": "/300x300.png", "phoneNumber": "(214) 210-4674", "address": { "street": "123 Main St", "city": "North Rodgerstad" }, "amenities": [ ... ] } (from 8KB to 312B -> 93% lighter response) -
Mobile Code – Initialize App Sync Client do { //
create app sync configuration (cache, service config etc) let config = … // client initialization let appSyncClient = try AWSAppSyncClient(appSyncConfig: config) } catch { fatalError("Error initializing appsync client. \(error)") } -
Mobile Code – Run a Query // call a GraphQL
Query appSyncClient?.fetch(query: GetHotelQuery(hotelId: hotelId), cachePolicy: .fetchIgnoringCacheData) {(result, error) in guard let r = result, error == nil else { print(error?.localizedDescription ?? "") return } // use the data in the GUI self.hotelDetails = r.data?.getHotel … } -
Reservations: Design What data is required? Can we use existing
data sources? How can we enable via AppSync? Challenges? -
Reservations: GraphQL Schema type Reservation { confirmationNumber: ID! hotel: Hotel!
guest: User! startDate: AWSDate! endDate: AWSDate! rate: Int! } type Query { ... reservations(guestId : ID!):[Reservation] } type Mutation { createReservation(input:ReservationInput) deleteReservation(confNum: ID!) } -
Reservations: Query Resolver - Input #set( $todayString = $util.time.nowISO8601().substring(0, 10)
) { "version": "2017-02-28", "operation": "Query", "query": { "expression": "guestId = :gId AND startDate > :startDate", "expressionValues": { ":gId": $util.dynamodb.toDynamoDBJson($ctx.identity.sub), ":startDate": $util.dynamodb.toDynamoDBJson($todayString) } } } -
Reservations: Result query { guestReservations { confirmationNumber hotel { hotelId
name image phoneNumber } startDate endDate } } -
Extending Reservation Workflow Legacy System /graphql Subscription (MQTT) Mutate reservation
record, modifying record in DynamoDB and triggering subscription DynamoDB Stream: Put event in queue for processing AWS Lambda AWS Lambda Amazon DynamoDB Amazon SNS Amazon SNS -
Mobile Code – Subscribe to changes let sub = CreateReservationEventSubscription(guestId:
userId) newReservationSubscription = try appSyncClient?.subscribe(subscription: sub, resultHandler: { (result, transaction, error) in // check for error // store a reference to the new booking into our cache transaction?.update(query: GuestReservationsQuery(guestId: userId)) { (data: inout GuestReservationsQuery.Data) in data.guestReservations?.append(bookingToAdd) // consume the data in the UI self.bookingDetails?.append(bookingToAdd) } } -
Adaptive Rates: Design R&D is building an adaptive rate capability
Delivered via Amazon SageMaker API How do we add today’s rate to our existing API? -
How to evolve our GraphQL API GraphQL supports multiple data
sources per query: 1. Add new field 2. Configure new Data Source and Resolvers -
Adaptive Rates: GraphQL Types and Resolvers type Hotel { hotelId:
ID! name: String! location: String! ... rate: Rate } type Rate { hotelId: ID! rate: Int! currency: String! date: AWSDate! } ## rate Field Resolver: Request Mapping ## { "version": "2018-05-29", "method": "GET", "resourcePath": "/rd/rates/${ctx.source.hotelId}", "params": { "headers": { "Content-Type": "application/json" } } } -
Adaptive Rates: Result query { getHotel(hotelId: 3) { name location
image phoneNumber address { street city } amenities rate { currency rate } } } { "name": "ElastiLodge North Rodgerstad", "location": "North Rodgerstad", "image": "https://placehold.it/300x300.png", "phoneNumber": "(214) 210-4674", "address": { "street": “...", "city": “..." }, "amenities": [ ... ], "rate": { "currency": "USD", "rate": 312 } } -
What did we accomplish today? Built a rich, flexible, mobile-ready
API Incorporated legacy, new, and prototype features in a single endpoint Evolved API as requirements changed No coding required! Well, there were a few Velocity Templates AWS Console includes starter / boilerplate templates -
AWS Amplify CLI Toolchain for developers Enables you to prototype
and build quickly Built-in GraphQL Transformer Converts schema to CloudFormation templates and builds cloud resources $ amplify api add # create GraphQL schema # for example: type Hotel @model @searchable { name: String! location: String! ... } github.com/aws-amplify/amplify-cli -
Backend Dev: Hmm. So you’re saying this “GraphQL” will allow
any web/mobile developer to arbitrarily query basically any field in any backend service, recursively, however they want, without any backend developers involved? Frontend Dev: Yeah, right? It’s amazing! […silence…] Backend Dev: Guards, seize this person. -
Alex Casalboni Technical Evangelist, AWS @alex_casalboni @ 2019, Amazon Web
Services, Inc. or its Affiliates. All rights reserved Thank you! bit.ly/graphql-aws-milano