Common mistakes with the October 2026 Gitea 1.23 surge

The migration and resource trap

I see teams fail when they ignore the massive hardware gap between lightweight forges and heavy DevOps platforms. GitLab Community Edition requires 16GB of RAM and 8 vCPUs for a single-node deployment. Software will not run on a 1GB VPS even if users attempt to run it on 8GB for memory-constrained environments. Gitea and Forgejo require much less, with Gitea needing only 512MB of RAM. I find that teams who attempt to migrate from Gitea to Forgejo often hit a wall when they realize that their existing databases do not support an in-place upgrade after certain version releases occur. This incompatibility forces a manual export and re-import of all repositories and data.

The Gitea 1.23 surge highlights these compatibility issues. While Gitea 1.23.0 provides updates, the underlying architectural split with Forgejo remains. Forgejo 17.0.0 released on 15 October 2026, and its 16.0 line support ends on 29 October 2026. If a team tries to move a Gitea instance directly into a Forgejo container without a clean migration, they will encounter broken databases. I would skip the attempt to bypass a clean install if you, as a sysadmin, value your data integrity.

I have seen migration attempts fail because users do not check their system notices. In Gitea, if a migration has a problem, users must check the logs or the System Notices in the admin interface. Users running Gitea in Docker often find that they must restart the binary to see any new logs. One user reported that a migration failed with a 404 error, yet the logs provided no specific reason. Gitea also includes a built-in package registry for npm, Maven, and PyPI.

Governance and legal constraints

The decision between Gitea and Forgejo hinges on legal constraints. Gitea uses the MIT license, which permits users to fork the code or embed it in commercial products without contacting the Gitea team. Forgejo uses the AGPL license, which requires anyone modifying and running the software as a network service to share those changes. Gitea Ltd manages the Gitea project, while the non-profit Codeberg e.V. governs Forgejo.

I see teams choose Forgejo when they want to avoid a commercial board making unilateral decisions. Forgejo development happens through public RFC-style discussions on Codeberg. Codeberg itself grew from 300,000 repositories in November 2025 to over 650,000 projects by June 2026. Gitea users often favor the larger ecosystem and the faster feature development speed. I would recommend Forgejo if a project needs a non-profit foundation to ensure long-term community control.

I find the Gitea commercial layer to be the primary driver for the Forgejo fork in late 2022.

Scaling and feature bloat

Scaling a self-hosted service involves managing both resource costs and feature bloat. GitLab CE bundles a full suite, including issue tracking, a container registry, and a CI/CD engine. This breadth comes with high infrastructure costs. A team of 40 engineers paying for GitLab.com Premium faces a bill that scales with headcount. Running Gitea or Forgejo on a single mid-tier VPS provides a fixed infrastructure cost that stays low whether the team has 10 or 100 members.

I see companies struggle with GitLab’s resource-heavy nature. GitLab requires many cooperating services, like Puma, Sidekiq, and Gitaly, to function. This complexity makes backups non-trivial because administrators must back up the repository, database, LFS, artifact, and registry storage consistently. The official GitLab-backup tool covers this, but administrators must test restores regularly. Gitea and Forgejo are much simpler, as they operate as a single Go binary plus a database.

I would advise against GitLab for small teams that only need simple Git hosting. Does a team with five developers really need the complexity of a full DevOps platform? For those users, Gitea or Gogs is a better fit. Gogs runs on as little as 64MB of RAM and provides basic repository management via SSH and HTTPS. If a team needs to connect Gitea to GitLab, they can use OAuth2 or OpenID Connect to unify login flows. This allows GitLab’s pipelines to run automatically whenever Gitea pushes new commits. One alternative is OneDev, which uses about 500 MB of RAM and runs pipeline steps inside Docker containers the server manages.