A couple of years have passed since Thoughtworks-funded development on GoCD ended. A number of folks have asked about GoCD's general project, maintenance, development and support status - along with some speculation. While some of this has been discussed in the GoCD forums over time, we thought it might help to add some clarity for users here.

We hope for this to give a fair summary of the current state so you can make the best choice for you or your organization.

The following represents the opinions solely of the writers (Chad Wilson & Aravind SV) in their personal capacity contributing to GoCD maintenance, and not an official position of Thoughtworks.

TL;DR

As of early 2023, GoCD has a very small open-source maintainer group who are still driving a limited amount of bug fixes, minor enhancements, periodic releases and otherwise supporting the wider community.

Since January 2021, GoCD has had no organizational backing behind development, nor a commercial/enterprise support offering.

Background

Here's a brief history of how we got here:

  • 2007: GoCD was originally developed as a commercial, closed-sourced product named Cruise by Thoughtworks Studios, as an evolution of CruiseControl, the first continuous integration server. In contrast to CruiseControl, GoCD's design instead focused on supporting continuous delivery with complex workflows, rather than purely supporting the build automation at the core of continuous integration practices.
  • 2010: Cruise was renamed to Go. (later slightly adjusted to GoCD to disambiguate from the increasingly popular golang language)
  • 2011: The Go Community Edition is made license-free, easing use for folks in many settings.
  • 2014: GoCD was open-sourced, with certain features (including support) remaining commercial closed source features requiring licensing from Thoughtworks.
  • November 2019: Thoughtworks announces the plans for full transfer of GoCD to the community and end of the commercial support offering.
  • May 2020: Previously commercial features/plugins are all open-sourced or capability directly incorporated into GoCD by the GoCD Core Team.
  • December 2020: Commercial support for GoCD ends, as does Thoughtworks-funded development and maintenance of GoCD.

How is GoCD currently being maintained?

As of February 2023, maintainer access to GoCD currently sits with a handful of folks that are either current or ex-Thoughtworks employees, including some of those originally part of the GoCD core team within Thoughtworks Studios. In practice, active contributions are coming from a much more limited set of folks - solely within their free time.

This very small contributor team has completed 3-4 relatively minor releases per year throughout 2021 and 2022, mainly focused around security-related updates, bug fixes and minor enhancements.

Does Thoughtworks continue to sponsor GoCD?

For now, Thoughtworks still graciously provides funding for the website & build infrastructure of GoCD, along with code signing certificates, and a single commercially licensed component required for Windows support. This is done on a goodwill basis, and there is no official agreement with or promise to the community regarding this support.

Does GoCD have active ongoing feature development?

At time of writing, there is no dedicated team actively contributing to GoCD feature development or major enhancements. There is also no ongoing funding behind open source contributions to GoCD, and GoCD has a very small contributor community.

In the current state, this makes it unlikely that GoCD will receive major enhancements or changes in the foreseeable future, unless there is a large further increase in contributions by users, organizations or other members of the community.

Is GoCD end-of-life? Abandoned? Should GoCD be considered deprecated?

GoCD is neither officially end-of-life, nor fully abandoned - but it's fair to say that its development is not thriving either.

As to deprecation, generally this is a decision that you should make within your specific context. If GoCD meets your particular needs and functions as you would like in your environment with its current feature set and user experience, and the risks are acceptable to you, you should feel free to continue to use it - as with any free & open source software.

What happened to moving GoCD to an open source foundation?

There was an attempt to move GoCD over to an open source foundation. Based on initial conversations, the Linux Foundation and the Apache Software Foundation were interested and would likely have accepted GoCD. However, that effort is currently dormant. We realized that given the small contributor base, the move to a foundation would not change the situation significantly. The potential challenge around funding the build and release infrastructure would remain. The legal effort to make that change would have not been in line with the value of that change. That effort could be restarted at a later time.

Is it risky to continue to use GoCD from a security standpoint?

Maybe? That depends on your assessment of risk. The security policy has some important caveats, but a balanced assessment of risk for open source software generally includes assessment of the vibrancy of the software's community.

Here are some additional aspects to consider:

The "good" news:

  • We believe GoCD was built with solid security principles in mind that have generally stood the test of time.
  • Software dependencies incorporated within GoCD are generally still under active development, and thus being patched/updated. These are being incorporated into new GoCD releases without major issue.
  • GoCD runs on and is built/test using modern, maintained language runtimes (Java 17 LTS, NodeJS 20 LTS, Ruby 3.1)
  • GoCD continues to have good compatibility with latest operating systems, which can be extended to its supported Docker images.

The "neutral" news:

  • With GoCD, we do not maintain/patch old releases of GoCD. In general, GoCD has always maintained robust backwards compatibility and sign-posted removals/deprecations in advance with an expectation that folks will move forward to the latest release.

The "not-so-good" news:

  • GoCD has a significant number of plugins, many different supported operating systems, many Docker image variants, a number of languages used across the codebase and many different features that make it difficult for a small team to support. Keeping on top of all of these requires non-trivial effort, and is likely not indefinitely sustainable with the current contributor/maintainer community size.
  • At time of writing (updated: December 2023), GoCD does depend on some libraries/frameworks which require significant effort to upgrade, but are end-of-life, and thus present some level of ongoing risk. If a major vulnerability is discovered in these libraries/frameworks that affects GoCD it will either require an urgent upgrade, or forking and patching of those libraries/frameworks. The community is not well placed to manage this kind of development. To help you assess risk, at time of writing, they include (but may not be limited to)

    • Spring Framework 4.3 (EOL December 2020)
    • Spring Security 4.2 (EOL October 2020)
    • Hibernate ORM 3.6 (EOL February 2012)
    • A small part of the UI relies on
      • AngularJS 1.0 - plugin view templating support (EOL November 2013)
      • jQuery 1.7 - stage and build/job detail views (subsequent upgraded in GoCD 23.4.0)
      • PrototypeJS 1.6 - stage and build/job detail views (subsequently replaced with jQuery in GoCD 23.5.0)

How can my organization or myself help with maintenance of GoCD?

You can learn how to contribute here and check out the resources available for getting started.

Should I migrate away from GoCD?

Paraphrasing from the original announcement – you should assess your needs and select the best tool for you. GoCD will be available as an open source project for use in the future. Thoughtworks and the GoCD maintainers will not be providing alternative tools recommendations or migration support. Different tools in this space offer different levels of modeling capabilities and features.

What's the best way to migrate away from GoCD?

GoCD has a reasonably comprehensive set of APIs you can use to export pipeline configuration as a starting point for partially automated migration. If you don't use pipelines as code techniques, you might want to start with using the APIs to get all pipeline groups and then get pipeline config for each pipeline for conversion to your target tool.

GoCD is great! But this doesn't seem like a great situation?

We think GoCD is great too. And we agree the maintenance situation is not great either. Come encourage us and help us out on the forums, or GitHub.