There is a separate page for our Product Vision for GitLab at the end of 2018.
This page describes the direction and roadmap for GitLab. It's organized from the short to the long term.
GitLab's direction is determined by GitLab the company, and the code that is sent by our contributors. We continually merge code to be released in the next version. Contributing is the best way to get a feature you want included.
On our issue tracker for CE and EE, many requests are made for features and changes to GitLab. Issues with the Accepting Merge Requests label are pre-approved as something we're willing to add to GitLab. Of course, before any code is merged it still has to meet our contribution acceptance criteria.
What our customers want
As a company, GitLab tries to make things that are useful for our customers as well as ourselves. After all, GitLab is one of the biggest users of GitLab. If a customer requests a feature, it carries extra weight. Due to our short release cycle, we can ship simple feature requests, such as an API extension, within one or two months.
On our releases page you can find an overview of the most important features of recent releases and links to the blog posts for each release.
GitLab releases a new version every single month on the 22nd. Note that we often move things around, do things that are not listed, and cancel things that are listed.
This page is always in draft, meaning some of the things here might not ever be in GitLab. New premium features are indicated with "Premium" label. This is our best estimate of what will be new premium features, but is in no way definitive.
The list is an outline of tentpole features – the most important features of upcoming releases – and doesn't include any contributions from volunteers outside the company. This is not an authoritative list of upcoming releases - it only reflects current milestones.
Starter features are available to anyone with an Enterprise Edition subscription (Starter, Premium, Ultimate).
Premium features will only be available to Premium (and Ultimate) subscribers.
- Reviews: batch comments on merge requests 11.3
- Instance level template repository 11.3
- Group-level project templates 11.3
- Timed incremental rollouts 11.4
- Show code quality notices on diffs/MRs 11.4
- Suggest approvers based on code owners 11.4
- Control incremental rollouts in the UI 11.4
- Implement inline code coverage remarks inside merge request changes tab file diffs 11.5
- Merge train/Release train/Merge when master succeeds: run build on merged code before merging 11.5
- Assign approvers based on code owners 11.5
- File template repository 11.5
- Create GitLab branch in JIRA issue in development panel 11.5
- Flaky test detection, reporting, prevention, and minimization 11.5
- Detect and report on flaky tests 11.5
- Multiple milestones per issue or merge request 11.5
- Allow users to create and manage feature flags for their applications (Feature Flags MVC) 11.5
- Testing for multiple browsers using Selenium 11.6
- Dashboard of all environments in a group 11.6
- Integrated Load Testing 11.6
- CI View for Selenium 11.6
- A/B testing of branches with GitLab Pages (Product)
- Make cross-repo CI triggers first-class Backlog
- Better than Atlassian Jira integration Backlog
- Code analytics Backlog
- Custom fields in issues Backlog
- Code coverage graphs and adding more metrics Next 4-7 releases
- Graph that shows code coverage on default branch over time Backlog
- Approvers based on code Git blame Backlog
- Multi project pipeline Backlog
- Add cross-project
dependenciesto .gitlab-ci.yml Next 4-7 releases
- Audit log improvements (META)
- Rebuild/update all containers
- Verified and reproducible builds
- Feature management (feature flags for user apps, not GitLab itself)
- Configuration Management Integration (Chef, Puppet, Ansible, Salt)
- Turnstile security: self learning and location dependent
- Watermarking of downloaded binaries
- Advanced compliance reporting (on access rights)
- [meta] Disaster Recovery Next 3-4 releases
- Label sets Backlog
- Expand upstream/downstream pipelines inline Backlog
- Import JIRA project issues into GitLab project issues Backlog
- Feature monitoring Backlog
- Advanced deploys (Blue/green, Canary, Traffic vectoring)
- Add "Jira integration" to multiple projects at once Backlog
- Group level integration with JIRA Backlog
- SLO and auto revert/stop incremental deployment Next 4-7 releases
- Block deploy/promote/merge if degrade performance too much
- More control over manual action permissions
- Controllable permissions for CI tokens
- Operator role
- Multiple blocking approver groups Next 7-13 releases
- Customized Performance Dashboards
- Automatically assign SLA time for Service Desk issues
- Automatically assign Service Desk issues to users
- Blocking issues Backlog
- Make GitLab CI/CD work with an external repository (GitHub, Bitbucket) Next 4-7 releases
- Automatically Load-balance tests Backlog
- GitLab PaaS Next 7-13 releases
- Create GitLab merge request in JIRA issue in development panel Backlog
- View GitLab merge request approvals information in associated JIRA issue Backlog
- Service Desk Analytics Backlog
- [meta] Enterprise Authentication & User Management
- Real-time updates in Jira dev panel Backlog
- Code Quality++ Backlog
- Scale deployments directly from the environment page Backlog
- Autoscaling apps Next 7-13 releases
- Application idling Next 7-13 releases
- Binary repository user interface Backlog
- Enforce file owners on protected branches Next 4-7 releases
- META: Code review flow which doesn't rely on manual assignment
- Increase visibility of incremental rollouts in deploy boards Backlog
- Compliance features Backlog
- Ensure we have a complete monitoring package for GitLab HA
- [Meta] GitLab HA improvements post GA Next 7-13 releases
Ultimate is for organisations that have a need to build secure, compliant software and that want to gain visibility of - and be able to influence - their entire organisation from a high level.
From a high level, the first major initiatives for Ultimate are:
Ultimate features will only be available to Ultimate subscribers.
Below are features that represent the various functional areas we see GitLab going in. This list is not prioritized. We invite everyone to join the discussion by clicking on the items that are of interest to you. Feel free to comment, vote up or down any issue or just follow the conversation. For GitLab sales, please add a link to the account in Salesforce.com that has expressed interest in a wishlist feature. We very much welcome contributions that implement any of these things.
Build and packaging
GitLab is the engine that powers many companies' software businesses so it is important to ensure it is as easy as possible to deploy, maintain, and stay up to date.
Today we have a mature and easy to use Omnibus based build system, which is the foundation for nearly all methods of deploying GitLab. It includes everything a customer needs to run GitLab all in a single package, and is great for installing on virtual machines or real hardware. We are committed to making our package easier to work with, highly available, as well as offering automated deployments on cloud providers like AWS.
We also want GitLab to be the best cloud native development tool, and offering a great cloud native deployment is a key part of that. We are focused on offering a flexible and scalable container based deployment on Kubernetes, by using enterprise grade Helm Charts.
GitLab High Availability
CI / CD
We want to help developers get their code into production; providing convenience and confidence to the developer in an integrated way. CI/CD focuses on steps 6 through 9 of our scope: Test (CI), part of Review (MR), Staging (CD), and part of Production (Chatops). When viewed through the CI/CD lens, we can group the scope into CI, CD, and things that are currently beyond any definition of CD.
We define our vision as “Auto DevOps”: leveraging our single application, it is simple to assist users in every phase of the development process, implementing automatic tasks that can be customized and refined to get the best fit for their needs. Our idea is that the future will have “auto CI” to compile and test software based on best practices for the most common languages and frameworks, “auto review” with the help of automatic analysis tools like Code Climate, “auto deploy” based on Review Apps and incremental rollouts on Kubernetes clusters, and “auto metrics” to collect statistical data from all the previous steps in order to guarantee performances and optimization of the whole process. Dependencies and artifacts will be first-class citizens in this world: everything must be fully reproducible at any given time, and fully connected as part of the great GitLab experience.
Many of the issues describe development of an n-tier web app, but could equally be applied to an iOS app, Ruby gem, static website, or other type of project.
See a slightly more complete rendering of an example pipeline.
GitLab CI provides an explicit
build stage and the concept of build artifacts, but we might need to separate out the build artifacts from test artifacts. For example, you might want your test runner to create a JUnit-style output file which is available for external consumption, but not included in the build image sent to production. Creation of an explicit build aligns well with Docker where the result of the build stage is a Docker image which is stored in a registry and later pulled for testing and deployment.
A key part of CD is being able to deploy. We currently have this ability via scripts in the
deploy stage in
.gitlab-ci.yml. We will go further.
GitLab is used to create, collaborate, review and manage content; often source code, by many businesses. We want to make it possible for everyone to be able to contribute content and feedback using GitLab.
Source code management
Inner sourcing, collaboration across organizational boundaries and contibuting patches to upstream open source projects means embracing the benefits of Git being a tool built for distributed software development. GitLab helps teams collaboratively write software, and will make it easier to collaborate across organizational and server boundaries.
A key part of writing and deploying high quality code is a thorough code review approval process. Merge requests provide this and are integrated with GitLab CI for testing, security testing, monitoring and more. They are a key component of GitLab's integrated Complete DevOps lifecycle.
Providing the tools for great code reviews improves code quality and helps teams iterate faster. We will help teams move faster by supporting more sophisticated commit, review, and merge strategies beyond treating a merge request as a single patch.
No current issues
Open Source Projects
Performance is a critical aspect of the user experience, and ensuring your application is responsive and available is everyone's responsibility. We want to help address this need for development teams, by integrating key performance analytics and feedback into the tool developers already use every day.
As part of our commitment to performance we are also deeply instrumenting GitLab itself, enabling our team to improve GitLab peformance and for customers to more easily manage their deployments.
Version Control for Everything
Moonshots are big hairy audacious goals that may take a long time to deliver.
Our vision is to replace disparate DevOps toolchains with a single integrated application that is pre-configured to work by default across the complete DevOps lifecycle. Consider viewing the presentation of our plan for 2018.
We try to prevent maintaining functionality that is language or platform specific because they slow down our ability to get results. Examples of how we handle it instead are:
- We don't make native mobile clients, we make sure our mobile web pages are great.
- We don't make native clients for desktop operating systems, people can use Tower and for example GitLab was the first to have merge conflict resolution in our web applications.
- For language translations we rely on the wider community.
- For Static Application Security Testing we rely on open source security scanners.
- For code navigation we're hesitant to introduce navigation improvements that only work for a subset of languages.
- For code quality we reuse Codeclimate Engines.
- For building and testing with Auto DevOps we use Heroku Buildpacks.
Outside our scope are:
- Network (fabric) Flannel, Openflow, VMware NSX, Cisco ACI
- Proxy (layer 7) Envoy, nginx, HAProxy, traefik
- Ingress (north/south) Contour, Ambassador,
- Service mesh (east/west) Istio, Linkerd
- Container Scheduler we mainly focus on Kubernetes, other container schedulers are: CloudFoundry, OpenStack, OpenShift, Mesos DCOS, Docker Swarm, Atlas/Terraform, Nomad, Deis, Convox, Flynn, Tutum, GiantSwarm, Rancher
- Package manager Helm, ksonnet
- Operating System Ubuntu, CentOS, RHEL, CoreOS, Alpine Linux
ML/AI at GitLab
Machine learning (ML) through neural networks is a really great tool to solve hard to define, dynamic problems. Right now, GitLab doesn't use any machine learning technologies, but we expect to use them in the near future for several types of problems:
Signal / noise separation
Signal detection is very hard in an noisy environment. GitLab plans to use ML to warn users of any signals that stand out against the background noise in several features:
- security scans, notifying the user of stand-out warnings or changes
- error rates and log output, allowing you to rollback / automatically rollback a change if the network notices abberant behavior
Automatically categorizing and labelling is risky. Modern models tend to overfit, e.g. resulting in issues with too many labels. However, similar models can be used very well in combination with human interaction in the form of recommendation engines.
- suggest labels to add to an issue / MR (one click to add)
- suggest a comment based on your behavior
- suggesting approvers for particular code
Because of their great ability to recognize patterns, neural networks are an excellent tool to help with scaling, and anticipating needs. In GitLab, we can imagine:
- auto scaling applications / CI based on past load performance
- prioritizing parallized builds based on the content of a change
Similar to DeepScan.
Similar to Sourcegraph.
Git was primarily designed for code, where it is dominating the world. Organisations working with large files, for instance art assets when creating films or games, are looking to get the power that GitLab offers, but can't adopt because of the little support for large files that Git offers.
GitLab is going to help these organisations adopt Git and migrate away from legacy platforms such as Perforce. There are a number of moving parts that make this possible:
Git LFS allows you to work well with large files in Git.
File locking is crucial to collaborating on non-mergeable files. Currently GitLab offers branch-limited file locking and Git LFS offers locking across branches.
Our goal is to make it faster and easier for groups of people (organisations, open source projects) to bring value to their customers. This is measureable:
- Faster: cycle time
- Easier: amount of steps to go from an idea to having it fully available (in production, monitored)
Plays well with others
We understand that not everyone will use GitLab for everything all the time, especially when first adopting GitLab. We want you to use more of GitLab because you love that part of GitLab. GitLab plays well with others, even when you use only one part of GitLab it should be a great experience.
GitLab ships with built-in integrations to many popular applications. We aspire to have the worlds best integrations for Slack, JIRA, and Jenkins.
Many other applications integrate with GitLab, and we are open to adding new integrations to our applications page. New integrations with GitLab can very in richness and complexity; from a simple webhook, and all the way to a Project Service.
GitLab welcomes and supports new integrations to be created to extend collaborations with other products. GitLab plays well with others by providing APIs for nearly anything you can do within GitLab. GitLab can be a provider of authentication for external applications. And of course GitLab is open source so people are very welcome to add anything that they are missing. If you are don't have time to contribute and am a customer we gladly work with you to design the API addition or integration you need.
Deployments should never be fire and forget. GitLab will give you immediate feedback on every deployment on any scale. This means that GitLab can tell you whether performance has improved on the application level, but also whether business metrics have changed.
Concretely, we can split up monitoring and feedback efforts within GitLab in three distinct areas: execution (cycle analytics), business and system feedback.
With the power of monitoring and an integrated approach, we have the ability to do amazing things within GitLab. GitLab will be able to automatically test commits and versions through feature flags and A/B testing.
Business feedback exists on different levels:
- Short term: how does a certain change perform? Choose A/B based on data.
- Medium term: did a particular new feature change conversions, engagement
Long term: how do larger efforts relate to changes in conversations, engagement, revenue
- A/B Testing of branches
You application should perform well after changes are made. GitLab will be able to see whether a change is causing errors or performance issues on application level. Think about:
- Response times of e.g. a backend API
- Error rates and occurrences of new bugs
- Changes in API calls
We can now go beyond CI and CD. GitLab will able to tell you whether a change improved performance or stability. Because it will have access to both historical data on performance and code, it can show you the impact of any particular change at any time.
System feedback happens over different time windows:
Execution Feedback & Cycle Analytics
GitLab is able to speed up cycle time for any project. To provide feedback on cycle time GitLab will continue to expand cycle analytics so that it not only shows you what is slow, it’ll help you speed up with concrete, clickable suggestions.
Why cycle time is important
The ability to monitor, visualize and improve upon cycle time (or: time to value) is fundamental to GitLab's product. A shorter cycle time will allow you to:
- respond to changing needs faster (i.e. skate to where the puck is going to be)
- ship smaller changes
- manage regressions, rollbacks, bugs better, because you're shipping smaller changes
- make more accurate predictions
- focus on improving customer experience, because you're able to respond to their needs faster
When we're adding new capabilities to GitLab, we tend to focus on things that will reduce the cycle time for our customers. This is why we choose convention over configuration and why we focus on automating the entire software development lifecycle.
All friction of setting up a new project and building the pipeline of tools you need to ship any kind of software should disappear when using GitLab.
GitLab comes in 4 editions:
- Core: This edition is aimed at solo developers or teams that do not need advanced enterprise features. It contains a complete stack with all the tools developers needs to ship software.
- Starter: This edition contains features that are more relevant for organizations that have more than 100 potential users. For example:
- features for managers (reports, management tools at the group level,…),
- features targeted at developers that have to work in multi-disciplinary teams (merge request approvals,…),
- integrations with external tools.
- Premium: This edition contains features that are more relevant for organizations that have more than 750 potential users. For example:
- features for instance administrators
- features for managers at the instance level (reporting, management tools, roles,…)
- features to help teams that are spread around the world
- features for people other than developers that help ship software (support, QA, legal,…)
- Ultimate: This edition contains features that are more relevant for organizations that have more than 5000 potential users. For example:
- compliances and certifications,
- change management.
Quarterly Objectives and Key Results (OKRs)
To make sure our goals are clearly defined and aligned throughout the organization, we make use of OKR's (Objective Key Results). Our quarterly Objectives and Key Results are publicly viewable.
From development teams to marketing organizations, everyone needs to collaborate on digital content. Content should be open to suggestions by a wide number of potential contributors. Open contribution can be achieved by using a mergeable file format and distributed version control. The vision of GitLab is to allow everyone to collaborate on all digital content so people can cooperate effectively and achieve better results, faster.
Ideas flow though many stages before they are realized. An idea originates in a chat discussion, an issue is created, it is planned in a sprint, coded in an IDE, committed to version control, tested by CI, code reviewed, deployed, monitored, and documented. Stitching all these stages of the DevOps lifecycle together can be done in many different ways. You can have a marketplace of proprietary apps from different suppliers or use a suite of products developed in isolation. We believe that a single application for the DevOps lifecycle based on convention over configuration offers a superior user experience. The advantage can be quoted from the Wikipedia page for convention over configuration: "decrease the number of decisions that developers need to make, gaining simplicity, and not necessarily losing flexibility". In GitLab you only have to specify unconventional aspects of your workflow. The happy path is frictionless from planning to monitoring.
We admire other convention over configuration tools like Ruby on Rails (that doctrine of which perfectly describe the value of integrated systems), Ember, and Heroku, and strive to offer the same advantages for a continuous delivery of software.