When comparing Eclipse Che vs Gitpod, the Slant community recommends Eclipse Che for most people. In the question“What are the best cloud IDEs?” Eclipse Che is ranked 8th while Gitpod is ranked 11th. The most important reason people chose Eclipse Che is:
Built-in terminal with root access so you can make changes to your running machines. Being able to SSH into the workspace so you can use a desktop IDE is handy.
Specs
Ranked in these QuestionsQuestion Ranking
Pros

Pro SSH + terminal
Built-in terminal with root access so you can make changes to your running machines. Being able to SSH into the workspace so you can use a desktop IDE is handy.

Pro Custom commands
You can package up custom commands with your workspace and then use them (or share them) with everyone else.

Pro Docker runtimes
You can choose from pre-configured environments for Java, Javascript, C++, PHP, C#, etc., or you can define your own by dropping in a Dockerfile - makes it easy for simple and complex projects.
Pro GIT and SVN VCS support
Projects can be easily imported from any Git or Svn repository hosting service.
Pro Reproducible environment

Pro Portable workspaces
The workspace in Che includes project sources, IDE and the runtime. So if you hand your Che workspace definition to another user and they execute it they will get everything they need to build, run and debug the project.
Also the runtime is in a Docker container so it will work even if the second user is on a different OS than the original user who shared their workspace with them.

Pro Previews
Che does a nice job to automatically map the service:port running in the Docker container (e.g. tomcat on 8080) to the Docker port it actually uses (something in the ephemeral range). You never need to figure that out - it's just made available when you run your server.
Pro Merge tool for VCS

Pro Open-source
Pro Customizable Workspaces
Gitpod workspaces are very customizable. You can define a workspace with .gitpod.yml
and define your own Dockerfile if you need tools that needs preinstalled automatically.
Pro Open Source to the core
Gitpod is a Web IDE product by TypeFox, which created Theia, a open source IDE framework. They rely on Kubernetes, Docker, Theia, and Let's Encrypt to provide this service - providing a 100% OSS-based SaaS.
Pro Workspaces are created from GitHub URLs
Any GitHub repository can be prepended with https://gitpod.io#
to define a Gitpod workspace. Here's an example of it.
Pro VNC Support
Gitpod has VNC support out of the box, which allows you to test Desktop apps within the cloud.
Pro Pre-authorized Git Client
This is the only GitHub-based Web IDE that pre-authorizes the git client, saving you time to do setup.
Pro GitHub Pull Request Issues Comments are viewable on the IDE
No need to jump tabs when making needed changes for your GitHub Pull Request. Gitpod supports viewing the comments and resolving those through the IDE - so you can resolve comments without ever needing to mark it resolved manually on GitHub.
Pro Automatic Configuration
If your GitHub repositories does not have a manifest to make a stack, Gitpod will attempt to make a stack based on what it understood about your codebase. However, for popular projects, a central repository on GitHub exists to configure a workspace automatically for them without the manifest file to be present.
Pro Language Server Protocol support
LSP is a standard made by Microsoft to create a single common core that can be consumed by a development tool. Theia is based on Visual Studio Code and supports LSP natively.
Cons
Con Slow runtime
Online IDE is much slower than desktop one.
Con Does not create forks automatically for you
If you make a Gitpod workspace based on a GitHub repository you do not own, git push
will fail and you need to query the Git client to make the fork for you.
Con Can be slow on 3G Networks
The crux of this is because of the large bundle.js
that is needed to be transferred, this is a problem within Theia.
Con Packages can't be installed through the CLI
Gitpod doesn't support apt install
s yet due to the container mounting through an NFS. To alleviate this, you need to define a custom stack for your project if you need it.
