Advantages of a single application

Single application

GitLab is a single application that does everything from planning to monitoring. The advantages of a single application for are listed in the following paragraphs. These result in more results, happier developers, improved cycle times, and a great return on investment.

Other vendors offer a kit plane you have to assemble yourself, GitLab is a type certified aircraft.

Single authentication

You only have to login to one application. Users don't waste time logging in. You only have to set up one account. It is clear to users what account they should use. Due to less confusion phishing is harder.

Single authorization

GitLab does not require you to manage authorizations for each of its tools. This means that you set permissions once and everyone in your organization has the right access to every component.

Single project

Users don't have to set up or ask others to set up a project across applications. This can be a major source of delays.

Single setup

With Auto DevOps your new projects have all the needed testing and deployment from the get go. And by having logging and security scanning happen automatically you reduce risk.

Single interface

A single interface for all tools means that GitLab can always present the relevant context. Relatedly, any feature is always in its most optimal location. We don't have to force ourselves to create a new page to serve a new feature. The feature will just appear in the best designed existing location, simplifying navigation.

Also, you don't lose information due to constant context switching between different interfaces.

Users don't have to switch between applications constantly. Furthermore, if you're comfortable with one part of GitLab, you're comfortable with all of GitLab, as it all builds on the same interface components.

Single installation

Running GitLab means that there are is only one single application to install, maintain, scale, backup, network, and secure.

Single upgrade

Updating GitLab means that everything is guaranteed to work as it did before. Maintaining separate components is often complicated by upgrades that change or break integration points, essentially breaking your software delivery pipeline. This will never happen with GitLab because everything is tested as an integrated whole.

Single data-store

GitLab uses a single data-store so you can get information about the whole software development lifecycle instead of parts of it. With multiple applications you not only have multiple databases but also different definitions, processes. Multiple data-stores lead to redundant and inconsistent data. You don't have to duplicate data nor do manual entry. There is a single source of truth without having to build a data warehouse.

Single overview

Automated links between environments, code, issues, and epics provide a better overview of project state and progress. All information is realtime and there is a coherent set of concepts and definitions.

Single flow

Using a single application means you don't have to integrate 10 different products. This saves time from your developers, reduces risk, increases reliability, and lowers the bill of external integrators. The flow is better since the more then 100,000 organization using GitLab use the same flow and more than 1800 people contribute to make it great. If you have a developer tool department they can now focus on other tasks to make your developers more effective.

Single vendor

You can deal with one vendor instead of multiple vendors pointing to each-other.

Single training

Your end user training becomes less complex; a multi-vendor environment means multiple trainers to manage.

Single codebase

We prefer to offer a single application instead of a network of services or offering plugins for the following reasons:

  1. We think a single application provides a better user experience than a modular approach, as detailed by this article from Stratechery.
  2. The open source nature of GitLab ensures that that we can combine great open source products.
  3. Everyone can contribute to create a feature set that is more complete than other tools. We'll focus on making all the parts work well together to create a better user experience.
  4. Because GitLab is open source the enhancements can become part of the codebase instead of being external. This ensures the automated tests for all functionality are continually run, ensuring that additions keep working. This is in contrast to externally maintained plugins that might not be updated.
  5. Having the enhancements as part of the codebase also ensures GitLab can continue to evolve with its additions instead of being bound to an API that is hard to change and that resists refactoring. Refactoring is essential to maintaining a codebase that is easy to contribute to.
  6. Many people use GitLab on-premises, for such situations it is much easier to install one tool than installing and integrating many tools.
  7. GitLab is used by many large organizations with complex purchasing processes, having to buy only one subscription simplifies their purchasing.

Emergent benefits of a single application

A single application across the complete DevOps lifecycle has unique, emergent benefits.

Some examples below.

CI and container registry

Pushing from CI to the container registry: before GitLab, this requires you to create a project and user account in a separate registry - and then on each job, credentials have to be passed between CI and the registry. With GitLab, none of this is necessary, because GitLab knows who you are and what your authorizations are in that project.

Debugging issues

Because all decisions, code, changes and deploys happen in the same place and are linked to issues and merge requests, GitLab has a full audit log of every decision, change and deploy. So if anything goes wrong, discovering the source of the issues is simple and doesn't require checking multiple apps, teams and logs.

Monitoring from merge request

You deploy a change by merging a merge request. Because GitLab has monitoring built-in, it's able to show in the merge request that e.g. the error rate has increased. The responsible developer sees this with the correct context and can start to solve the issue immediately.

Further examples

  1. Planning (code merge closes issues updates epic) - Always up to date and across different projects
  2. Move from issue board to issue tracker - Kanban boards are consistent with issue tracker
  3. Getting the stack running for a new developer - Can start developing without setup time or help
  4. Merge request contains monitoring, security, etc. information - Can see overview in MR of tests, quality, performance, browser performance, artifacts, discussion, environment, rollback
  5. Security (security in parallel, instead of separate) - Security tests results can be aggregated
  6. Identify vulnerable applications quickly and automate the remedy
  7. Add the build application to a container registry - No need to set up a project in the container registry and pass credentials around
  8. Need to find an artifact from a CI build - Artifacts are accessible directly from any merge request
  9. Reviewing the application - QA test can start work immediately
  10. Run CI/CD/Security/Quality without having to configure anything - One click to deploy to production
  11. Need to scale up a containerized application - Scaling up and down is just another button inside your project with all relevant data to make a decision
  12. A project adopts containerization - New servers are automatically provisioned via Kubernetes
  13. Rollback a failing deployment automatically without human intervention
  14. Activity feeds span all product categories, across DevOps lifecycle
  15. Cycle time (real time and integrated) - Can view activity across DevOps lifecycle
  16. Permission levels and settings - User rights are consistent across all applications
  17. Audit log - What a person was able to do across the lifecycle in audit logs
  18. Geo does not just to SCM, it also does CI, CD, issue tracking - People at other location have quick access to the whole DevOps-lifecycle, not just repos
  19. Easier to administer then 10 different applications - The flow between the activities doesn't break because of updates to a single step in the life-cycle
  20. Start a new microservice project - A project has all tools in it, no need to find out how this project is named in other tools
  21. Someone new joins the organization - Users automatically have access to all tools that are relevant for their role, without creating a ticket and waiting

Conversational Development

Conversational development carries a conversation across functional groups through the DevOps lifecycle, involving gatekeepers at every step. By providing relevant context, a feature that is only possible with an integrated solution like GitLab, we can reduce cycle time, making it easier to diagnose problems and make decisions.

Concretely, conversational development in GitLab means that a change can be easily followed from inception to the changes it made in performance and business metrics and feeding this information back to all stakeholders immediately.

Effectively, this allows cross-functional teams to collaborate effectively.

Review Apps

Review apps are the future of change review. They allow you to review not just the code, but the actual changes in a live environment. This means one no longer has to check out code locally to verify changes in a development environment, but you simply click on the environment link and try things out.

Only a tool that combines code review with CI/CD pipelines and integration with container schedulers (Kubernetes) is able to quickly create and shut down review apps and make them part of the review process.

Cycle Analytics

Cycle analytics tell you how what your time to value, from planning to monitoring is. Only by having direct access to each step in the software development lifecyle, GitLab can give actionable data on time to value.

This means you can see where you are lagging and make meaningful change to ship faster.

Container registry integrates with CI

Every GitLab projects comes with a container registry. That means there is no need for elaborate configuration to be able to use and push container images in CI. Rather, all you have to do is use a pre-defined variable in your CI configuration file (.gitlab-ci.yml).

Use cases, not modules

Sometimes users ask if GitLab has particular module that contains a set of features. A modules-mindset results in narrowly-focused solutions and redundant features and abstractions. GitLab is a single application that solves entire use cases that might touch many parts of GitLab, without creating additional unnecessary complexity.

Existing features and data accrue value over time

As GitLab breadth expands, but is maintained as a single application, new features are continually back-integrated into existing features. This means that existing features and the data that they have already generated over time accrue new value for very little cost in the context of the new features and their use cases.