September 2023 Summaries

This is a story about how a company, Railway, rebuilt its logging infrastructure to handle 1 billion new logs per day. The old system was struggling and they wanted a scalable solution that could support filtering and other advanced features. They discovered ClickHouse, a column-oriented database used for log aggregation and analytics, which seemed like an ideal fit. After testing and implementing ClickHouse, Railway was able to seamlessly integrate it into their existing logging architecture, allowing them to add new features like Log Explorer, which provides an environment-level view of all logs. The new system has improved performance, scalability, and speed, making it easier for the company to identify and fix problems quickly.

Railway has introduced a new feature called "Scale to Zero" which allows users to automatically stop and start workloads with no traffic, including databases, with the flip of a switch. This feature is made possible by leveraging Envoy proxies, namespaces, nftables, and cAdvisor to monitor and manage traffic. When a container is inactive for 5 minutes, it will be flagged as such and stopped, but its port will remain active, allowing traffic to be redirected to the container when it becomes active again. This feature aims to reduce downtime and improve scalability for users, with benefits ranging from cost savings for indiehackers to improved efficiency for Fortune 500 companies. The feature is now available in Priority Boarding, and users can toggle the "App Sleeping" setting to enable or disable it.

Railway, an infrastructure company, has released Changesets, its first major Lifecycle Management release, which enables versioning and reproducibility of infrastructure environments. This feature was inspired by Git and aims to simplify the process of managing infrastructure deployments, allowing developers to fork and merge environments with ease. The team used Mermaid diagrams to visualize the implementation and wrote tests that could be easily understood and debugged using a similar diagramming tool. With Changesets, Railway aims to support teams in statefully managing configuration as a team and plans to make its Config as Code and Changesets features fully versionable in the future.

So You Think You Can Scale?` Railway Launch Week 01 theme is Scale, a new infrastructure company providing instant application deployments to developers. The traditional vertical scaling approach can be overwhelmed by massive requests, causing stress and frustration. However, Railway's new horizontal scaling feature allows for easy scaling of services with just one dial setting. This feature includes automatic load balancing over both public and private networks, ensuring seamless communication between replicas. Additionally, the feature provides fault tolerance and replicates containers across multiple machines in Railway's fleet, making it a great tool for worker pools and distributed computing frameworks. The horizontal scaling unlocks parallel processing capabilities as well.

Railway, an infrastructure company that provides instant application deployments to developers, has shifted its focus from plugins to services for deploying databases. This change was necessary due to the complexity and scalability issues with maintaining custom plugins for each database. The new "Next-Gen Databases" feature allows users to deploy any type of database on Railway using a service-based approach, reducing the need for plugins. The feature provides several key primitives, including Volumes for persistent storage, Docker image deploys, TCP proxies for routing traffic, and template variable functions for generating secrets and passwords. These primitives enable users to create reusable templates for deploying databases in 1-click, making it possible to deploy any type of database globally on Railway.

Hello World: We’re Railway` is about Railway Launch Week 01, where the theme is Scale. Railway is an infrastructure company that provides instant application deployments to developers. They are expanding across three new regions: US-East, EU-West, and Southeast Asia, with plans for more in the future. The Regions feature allows users to easily switch between these regions without downtime, thanks to a private networking system and edge networks. This enables seamless traffic routing and load balancing. Railway's architecture is built on top of Temporal, which manages builds, deployments, routing, and cleanup. With Regions, users can now easily switch between regions with zero downtime. The company plans to expand its capabilities further, including hosting on different clouds, baremetal hardware, or user-provided infrastructure.

We're Railway V2, an infrastructure company that provides instant application deployments to developers. Our mission was to rethink the build, deploy, and observability stack for the modern web, aiming to make software development better. We started with a simple database deployment in 2021 and scaled to 100k+ users by year's end. However, as we grew, we faced challenges and self-doubt, questioning whether our solutions would scale and if Kubernetes was the solution we needed all along. After a disastrous attempt at building "Railway on Kubernetes," we took a step back, regrouped, and refocused on our core experience. We spent months rebuilding our infrastructure, fixing issues, and shipping new features, culminating in the launch of regions, our most requested feature. Our platform is now rock solid and feature complete, enabling users to ship anything on it with ease.