January 2020 Summaries

The text discusses the pros and cons of updatable views in PostgreSQL. Updatable views are a feature that allows users to insert or update data through a view instead of directly on the table. They were used by Hasura, an open-source engine that sits on top of Postgres, to enforce permissions for non-admin users. However, updatable views have certain limitations such as not having default column values and not supporting ON CONFLICT clauses. To address these issues, Hasura moved away from using updatable views and instead used a simpler approach involving insert CTEs and a helper function to check permissions.

The Hasura GraphQL Engine is designed for high performance and scalability. It can process a large number of queries (up to 1000 q/sec) with low latency, even in a small footprint of just 50MB RAM. Additionally, it supports massive concurrency, which is particularly useful for real-time applications. The performance metrics include high throughput and concurrency, as well as efficient handling of large queries and results. Hasura's GraphQL query processing architecture involves multiple stages, including parsing, validation, and planning. It compiles GraphQL to SQL using a compiler-based approach that allows for the formation of a single SQL query for any depth of GraphQL query. PostgreSQL prepared statements are leveraged to improve performance by skipping parsing in frequently repeated requests. This is particularly useful when dealing with authorization, as Hasura automates this process by providing developers with a fine-grained declarative auth DSL for every Postgres entity. JSON aggregations help in efficiently generating the SQL response into a neat JSON that the client can understand, reducing processing time and improving performance. Hasura's query caching eliminates parsing/validation for GraphQL queries, while prepared statements eliminate the same for PostgreSQL queries, resulting in high performance. Overall, Hasura's architecture ensures efficient handling of GraphQL queries and SQL execution plans, leading to improved performance and scalability.

Hasura is an open-source GraphQL server that provides instant realtime APIs over Postgres. It helps developers build GraphQL apps backed by Postgres or incrementally move to GraphQL for existing applications using Postgres. Key features include powerful queries, realtime capabilities, remote schema merging, webhook triggers, access control, high performance, and an admin UI with migrations. The Hasura GraphQL Engine can be deployed on various platforms such as Heroku, Docker, or DigitalOcean. It supports Postgres types and integrates with existing auth systems for field-level access control.

The second edition of GraphQL Asia will take place from February 20th to 22nd in Bengaluru, India. The conference will feature a dedicated day of workshops on the first day at Hasura's Terrace and two days of talks at Leela Palace hotel. Over 30 speakers from around the world will share their experiences, best practices, and insights about working with GraphQL. Topics include the future of GraphQL, using GraphQL in organizations, improving skills in GraphQL, and more. The conference will also offer four workshops ranging from beginner to advanced levels for attendees to get hands-on experience with GraphQL and related tools. Attendees can purchase passes for the workshop and conference at a discounted rate if they buy tickets in bulk.

This post discusses how to create an offline-first web app using RxDB and Hasura. The main goal is to make data changes locally first and then sync them with Postgres securely. RxDB, a NoSQL client-side database, is used for local storage, while Hasura provides a GraphQL server for syncing data. The post covers the setup of both tools, creating a simple schema, making the app functional offline, setting up replication using GraphQL, and handling conflict resolution and eventual consistency issues. It also mentions some challenges in building an offline-first app and recommends further reading on similar topics.