September 2023 Summaries

Henry Meds, a subscription healthcare service, has successfully scaled rapidly using Hasura. The talk "How Henry Meds Scaled Rapidly with Hasura" was presented by Nathaniel Armer, CTO and Founder of Henry. Hasura has improved developer productivity at Henry, enabling them to build scalable architectures and handle security in their fast-growing telehealth startup. Using Hasura, Henry shifted from building APIs to modeling their data, which significantly reduced the amount of backend engineering work they had to do. The use of built-in migrations, triggers for business logic, remote schemas to expose external endpoints, and turning actions into data have all contributed to Henry's successful scaling with Hasura.

Hasura conducted an internal two-day hackathon to validate and test features of its upcoming v3 release, which includes significant architecture changes and new features. The event aimed to create awareness within the company about Data Delivery Network (DDN), encourage innovation, and gather feedback for a successful public launch. Participants were encouraged to explore DDN's new concepts and build connectors using TypeScript. Seven projects were submitted, with winning ideas being granted the opportunity to see their concepts implemented. The hackathon helped validate Hasura DDN, identify fixes and improvements, and create internal awareness about the new product offering.

The Hasura MongoDB 1.0 connector has been released to provide users with cutting-edge database solutions and improve performance. Key tactics for achieving lightning-quick performance include database push-downs, query compiling, and reducing unnecessary data fetching and processing on the server's or clients' side. The Hasura MongoDB connector efficiently handles cross-collection relationships using lookup queries and incorporates access controls and permissions into its efficient framework. It also seamlessly integrates auto-generated query modeling for complex queries with multiple arguments, ensuring efficiency and optimized performance.

Hasura has introduced a new feature that allows users to deploy their own data connectors to Hasura Cloud using the Hasura Connector (Beta) CLI plugin. This plugin can be installed alongside the Hasura CLI and is used to deploy and manage Data Connectors on the same Hasura Cloud platform running GraphQL APIs. Deployments can take any public GitHub repo with a Dockerfile, build it, and deploy it on Hasura Cloud. The feature is currently in Beta and free of cost, but there are limitations on the number of connector deployments allowed per user based on their project tier.

In the August Community Call event, new features were launched for Hasura GraphQL Engine. These include MongoDB support with a 1.0 milestone, enabling users to create and update their GraphQL schema from NoSQL documents, join efficiently between collections, and enforce access control permissions on data. GenAI applications with Hasura Notebooks allow quick bootstrapping of AI applications by connecting favorite vector databases, using Hasura Event Triggers for auto-vectorizing data, and providing controlled data access to LLMs. A new CLI plugin for GraphQL federation CI/CD allows verification of metadata consistency before publishing changes to a supergraph. Remote table authorization enables combining data from different databases with authorized data. In August, three major releases were made: v2.31.0, v2.32.0, and v2.33.0, which introduced new features like suggested relationships UI, SQL Server stored procedures support, and ClickHouse support (Beta). The next Community Call Launch Day is scheduled for September 28th.

The text discusses the evolution of backend architectures and how modern technologies like React Server Components have sparked interest in these patterns. It highlights the use of data-bus like patterns for supporting multiple frontend applications or Backend for Frontends (BFF) by building out a Backend for Data (BFD), or capability based API. GraphQL is also mentioned as being well suited for data integration tasks. The text emphasizes that complexity is the enemy and that decoupled systems can help scale complexity and teams. It concludes by stating that understanding tradeoffs of each pattern or technology is crucial, and that a future post will delve into frontend pressures leading to different decisions.