May 2021 Summaries

Mark Ericksen discusses the use of Livebook, an Elixir project similar to Jupyter notebooks, for documenting and exploring critical pieces of code in applications. He demonstrates how it can be used with a TicTac game application running on Fly.io. The key features of Livebook include code completion, Elixir code execution, markdown text for explanations, keyboard shortcuts for easier navigation and editing, and storage of notebook files as plain Markdown files. These features make it easy to experiment with code, execute it, and demonstrate the central idea behind an application. Additionally, Livebook supports collaborative editing and can be connected to existing Elixir projects.

Fly now includes built-in Prometheus metrics for all its apps, allowing users to monitor performance, create alerts, and export their own custom metrics. The platform uses a Prometheus-style system with three types of metrics: counters, gauges, and histograms. Users can query these metrics using the PromQL language and visualize them in Grafana. Fly also supports custom metric generation for applications built on various frameworks such as Go, Python Flask, and Node.js.

Fly.io has developed a system that transforms container images into fleets of micro-VMs running around the world on their hardware. They have built an infrastructure for collecting, storing, and analyzing metrics to monitor the performance of these VMs. The metrics are collected using Prometheus exporters and stored in Victoria Metrics, a time-series database. Telegraf is used as a swiss-army knife tool that adapts arbitrary inputs to arbitrary output formats. Consul manages the configuration for all these components. This system allows users to monitor their applications running on Fly.io by exporting metrics and adding them to the fly.toml file. The metrics are then collected, stored, and displayed in Grafana Cloud, which provides a user-friendly interface for analyzing performance data.