BrikL - A GraphQL Native

BrikL's GraphQL journey started in 2016 with constantly changing business requirements towards finding the product market fit. The talk will share some insights on the journey and lessons learnt of a startup using GraphQL to enable high development velocity and how BrikL is using GraphQL today.

Transcript

1. BrikL A GraphQL Native Tobias Meixner Co-Founder & CTO Brikl Ltd - www.brikl.io 12 April 2019
2. Agenda 2 ▪ Startup Life ▪ BrikL GraphQL Journey ▪ Our GraphQL Stack
3. Tobias Meixner 3 ▪ Born in Berlin (born before it was cool) ▪ In Asia for over 6 years ▪ Based in Bangkok & Guangzhou ▪ Co-Founder & CTO @ ▪ @MeixnerTobias
4. Startup Life Developing software on fire
6. 6 Actually it is ... most of the time...
7. What is a startup? 7 ▪ Focus on growth ▪ Solve a problem ▪ More questions than answers ▪ Find product/market fit
8. 8 Startup Valley of Death Source: Osawa and Miyazaki, 2006
9. 9 Source: https://www.leadingagile.com/2017/11/planning-technical-debt/ Leverage ▪ Incur intentionally for a purpose ▪ Useful when cost-effective Waste ▪ Sloppy coding ▪ Refactoring has no immediate benefit Technical Debt
10. Dev in Startup... 10 ▪ Move fast, break things? ▪ Delete more than you add? ▪ Crave for stability and a constant?
11. BrikL constants since 2016 11
12. BrikL & GraphQL Journey with GraphQL as constant
13. About BrikL 13 ▪ 2D & 3D Design studio ▪ E-commerce Platform for Custom-Made Products ▪ www.brikl.io
14. Why GraphQL 14 ▪ Pure focus on Frontend development ▫ MVP-style / UI-Driven development ▪ GraphQL Schema as major constant in our development ▪ GraphQL Schema as discussion base between project shareholders
15. MVP (2016) 15 ▪ React ▪ Redux ▪ graphql-js ▪ AWS Lambda (NodeJS) ▪ Amazon DynamoDB Based on: serverless/serverless-graphql
16. v1 (2017) 16 ▪ React ▪ Redux Apollo Client ▪ graphql-js Apollo Server ▪ AWS Lambda (NodeJS) ▪ Amazon DynamoDB
17. BrikL GraphQL Stack “Serverless GraphQL”
18. 2018 - Today 18 ▪ React ▪ Apollo Client ▪ Apollo Server ▪ AWS Lambda (NodeJS) ▪ Amazon DynamoDB ▪ + Gatsby ▪ + Apollo Engine ▪ + Elasticsearch ▪ + Tooling: ▫ Serverless framework ▫ Cypress ▫ Sentry ▫ Netlify + CircleCI ▫ AWS stuff
19. Architecture 19 ▪ Serverless from Day 1 ▪ Managed Services where possible ▪ GraphQL as thin layer ▪ DynamoDB to ▫ Store “State” ▫ Emit events to streams ▫ Similar to Hasura 3factor app
20. Architecture 20 CDN Engine Server DynamoDB S3 Glue Elasticsearch ETL Stream Cognito JWT Authn/z X-Ray Logs Client Gatsby API Real time
21. Gatsby 21 ▪ Source data from GraphQL ▪ Generate storefront pages ▪ Source data based on selected store integrations ▫ Instagram, … ▪ Future > Content Mesh
22. Apollo Client 22 ▪ Query, Mutation components ▪ In-Memory-Cache ▪ Link Context > JWT ▪ Error link > Sentry ▪ HTTP Link > CDN vs POST ▪ SSR in Gatsby
23. Apollo Server 23 ▪ Apollo Server Lambda ▪ Apollo Engine ▪ Context > User, Config ▪ Cache Control > CDN ▪ graphql-shield > Authz ▪ graphql-tools > make Schema
24. Schema directives 24 ▪ Authorization ▪ Formatting ▪ Validation ▪ Data sourcing & ORM @makeSthAwesome
25. Schema-first 25 ▪ Great fit for getting started ▪ High velocity for startups ▪ Schema directives to rescue ▫ Schema = ORM ▫ SchemaDirectiveVisitor
26. Schema directives @ BrikL 26 ▪ S3 ▪ DynamoDB ▪ Elasticsearch
27. S3 directive 27
28. DynamoDB directive 28
29. What’s next? 29 ▪ Typescript / Code-first? ▪ State Management ▪ publish DynamoDB Datasource library ▪ Subscriptions...
30. Thank you Questions? Find me at the speaker hours ~5pm See you in