July 2023 Summaries

The text discusses strategies for ensuring redundant logs in case of network failures or other unexpected events. It suggests two approaches: logging to multiple places and using a custom log shipper. Logging to multiple places involves having the primary application write to a separate log file in addition to standard out, while using a custom log shipper entails subscribing directly to NATS and processing log entries yourself. Both strategies aim to provide additional layers of protection for crucial log data.

Fly.io has developed a multipurpose secret-using service called the Tokenizer, which is a stateless HTTP proxy that holds the private key of a Curve25519 keypair. When they get a new 3rd party API secret, they encrypt it to Tokenizer's public key; they "tokenize" it. Their API server can handle the (encrypted) tokenized secret, but it can't read or use it directly. Only Tokenizer can. They also developed SSOkenizer, which performs the OAuth2 dance on behalf of Rails and then uses the output to drive the Tokenizer. These tools are open-source and easy for others to deploy and use themselves.

Fly.io is preparing to launch Bun 1.0 on September 7th. As Fly.io has previously shown its love for JavaScript (JS), it initially started with Node.js support and is now expanding to other runtimes, beginning with Bun due to its aim for complete Node.js API compatibility. With flyctl version 0.1.54 and @flydotio/dockerfile version 0.3.3, users can launch and deploy bun applications using fly launch and fly deploy, provided they have installed the latest version of Bun (0.5.3 or later) and meet certain conditions in their package.json file. Fly.io also provides a sample application that can be deployed. However, it is important to note that everything is still in beta phase, and some issues were encountered during preparation for this support. Despite these challenges, fixes made for either Node.js or Bun framework will generally benefit both since they share the same generator.

Fly.io introduces LiteFS Cloud, a managed backup and point-in-time restore service for SQLite databases running on their platform or elsewhere. LiteFS is an open source distributed SQLite database that replicates data across nodes in an app cluster while maintaining the simplicity and speed of vanilla SQLite. However, it lacked disaster recovery features until now. LiteFS Cloud provides automated backups every five minutes (8760 snapshots per month) without requiring users to build their own backup systems. The service is currently in preview and free to use.