A new platform for documentation and tutorials is launching soon.
We are migrating Waypoint documentation into HashiCorp Developer, our new developer experience.
Waypoint is a new project and we have a lot we want to achieve with it. This page will outline some of the ideas we have for future versions of Waypoint. This is not an exhaustive list nor is it a promise that these features will be built, but it gives a glimpse into the vision we have for this project.
For a finer-grained view into our roadmap and what is being worked on for a release, please refer to the GitHub Issue Tracker. This is a more manual approach to reading our roadmap and includes small details such as bug fixes, but is much more accurate to what we're working on.
Waypoint has always had
release lifecycle events, but
to handle more complex deployments, Waypoint needed the ability to add custom steps.
We recently released a Tech Preview of custom pipelines,
introducing the ability to script custom steps, such as running tests after a
build or security scanning artifacts prior to a deploy. As we work to bring
custom pipelines out of Tech Preview, we'll be introducing more support for
steps consuming and producing data for other steps in a pipeline, improvements
around how pipelines access VCS data, customizable step plugins, and introducing
pipelines to the UI.
Waypoint has the workspaces feature for isolating different environments. We plan to build on this feature, and custom pipelines, to enable workspaces that may target different infrastructure environments (such as a dev vs. production cluster), and we want to build features so applications can be easily "promoted" from one stage to another without rebuilding artifacts in each workspace.
We plan to incorporate a tighter integration with Terraform to enable folks to provision infrastructure right from their waypoint.hcl file. Waypoint will be able to trigger Terraform to provision deployment infrastructure, and connect that infrastructure to the Waypoint server.
»Service Brokering (Databases, Queues, etc.)
We plan on having some mechanism by which applications can request services such as databases, queues, and so on. These services must be satisfied by some external system such as Amazon RDS. Waypoint will be able to request the service, connect the configuration, and destroy it when its no longer in use.
A design goal of Waypoint is explicitly to not do automated infrastructure management, so this feature will still put the burden of managing these services onto the user. However, Waypoint can help orchestrate initializing them and configuring applications.
We will likely integrate with Terraform in some way.
Waypoint currently associates a waypoint.hcl configuration with a single project. In the future, we will allow parameterized configurations to be shared as a template library with your installation. When users create a new project, they can select from a list of common templates.
The initial Waypoint release has a minimal token system that is used as a blanket authorization mechanism. In the future we will add a more fine-grained access control mechanism to Waypoint.
One exception to this today are Entrypoint tokens. The Entrypoint uses a special fine-grained token that allows it to only access entrypoint-related functionality for the specific deployment it represents. This was built as a special security mechanism however and isn't generally available to limit access for other tokens.
We're really excited about
waypoint exec but there
are a lot of improvements we want to make here:
New deployment targeting. - We want exec to be able to spin up new exec-focused deployments so that your exec sessions aren't shared with deployments that might be receiving user traffic. Additionally, this will let us harden security by disabling exec in most deployments.
Better security, specifically around disabling. - We want exec to be safe. One mechanism towards that is having exec be disabled as often as possible but can still easily be used when needed. We think having exec disabled by default paired with the above deployment targeting feature makes this a reality.
Native exec. - If a platform provides a native
execfeature, we want to provide that as an option to users. This is always likely to not be a default because we want the consistent option as the default.