Capabilities

Atomist is all about helping you ship high-quality software in less time. Whether you are working with large legacy codebases, your own “majestic monoliths”, or aiming for “zero-overhead microservices”, Atomist provides you with the tools to speed development, smooth operations, and close the code/runtime feedback loop.

Bringing everything together

Atomist consumes events from the platforms and tools you use every day. By linking related events from source code repositories, continuous integration, testing, and runtime environments, Atomist can present them in a unified manner to appropriate team members via direct or group messages. The result is massive reductions in context-switching costs and and shorter incident time-to-resolution.

Rug ties everything together

Rug is the programming model and runtime that ties everything together. Rug not only understands individual events, but can learn relationships between events and respond to events with appropriate, automated actions.

Development lifecycle automation

Atomist helps you develop better software more quickly by greatly enhancing your ability to automate your software development lifecycle. By intelligently linking your source code repository, your CI system, your deployment platform, your logs, and your alerting system, Atomist connects code changes to runtime changes and provides that information back, with its full context, back to the appropriate team member. It can can even automate remediation, obviating the need to alert anyone.

The Rug programming model was created to help you streamline and better manage your software delivery lifecycle. By writing Rugs, i.e., programs that implement the Rug programming model, you are able to automate code modifications and API calls up through complex workflows and decision points. There are several different types of Rugs, each providing a different way to help you improve your development and operations lifecycle.

Event-driven actions: Event Handlers
Rug event handlers respond to events. Event handlers can use the information on an event and its related events to message the right team members, take direct action, and/or create new events, which can trigger other event handlers. Need to notify someone when their commit causes a stack trace in production? Automate that with an event handler.
ChatDev and ChatOps: Command Handlers
Rug command handlers respond to messages in chat, allowing you quickly create bot commands to automate common tasks in software development and operation. Want to automate the release of a new version of your service to production? Write a command handler so you can initiate the release right from chat.
Project updates: Editors
Rug editors modify code directly and consistently. They can modify any code, in any project, regardless of whether the project was created by Atomist or not. Need to add or update a dependency and add some boilerplate code for it? Codify it in an editor.
Project creation: Generators
Rug generators create new projects from a model project. More powerful than templating approaches, generators are real projects operating under native tooling. Need to stamp out a bunch of new microservice projects? Stop the copy/paste/find/replace cycle and use a generator.

Individual Rugs can be composed. For example, editors can be used by Rug handlers, commands, and generators to modify code in a consistent manner across many repositories. Handlers can execute commands. Events can fire reviewers, e.g., every PR can trigger a reviewer that ensures it complies with your coding and submission standards.

Events and code selection

How does a handler know what event it should act on? How does an editor know what part of what file to edit? Atomist provides a powerful mechanism for pinpoint selection of events and code: path expressions. Inspired by XPath, path expressions provide a concise abstraction for navigating the various elements in your code and development environment. Path expressions select related events satisfying specific criteria, e.g., a Slack user whose GitHub user authored a commit that was in a push that triggered a CI build that failed, and specific blocks of code in specific files, e.g., all Scala methods that return Boolean.

Developing Rugs

Atomist is extensible. Atomist ships with many useful capabilities, but the ways in which Atomist can improve your development and operations do not stop there. The primary supported development language for writing Rugs is TypeScript, although any language that compiles to JavaScript can be used. You can modify existing Rugs or create your own Rugs to match your processes and workflows.

Atomist meets you where you are. It doesn’t make you change how your team develops and ships code. It improves your code and speeds the shipping of great software.