GHES Cluster High Availability

[github-product-roadmap]

Summary

In an effort to make it easier for customers to administer GHES and to minimize unintended mis-configurations, we will support a Cluster HA offering. This will enable a Hot/Hot configuration for stateless services and Hot/Warm cluster topology for stateful services/infrastructure.

Intended Outcome

Our Cluster HA solution will enable customers with the ability to scale horizontally and more efficiently manage traffic. We choose to be vigilant in meeting the evolving needs and expectations of our customers as their data and organizations grow on GitHub.

How will it work?

• Make use of secondary cluster hardware to serve traffic
• Fix cluster replication related bugs
• Cluster HA active/passive for stateful nodes only
• Cluster HA active/active (read-only)
• robust tests for Cluster HA
• Improved monitoring and observability

[ankneis]

This issue is being closed as outdated. For more information, please check out this Discussion post. Stay tuned for new additions to our refreshed public roadmap!

[ankneis]

We wanted to provide more details on why we removed this from the roadmap. Continuously improving GHES availability and operations at scale are important objectives for GitHub's product team. We are currently investigating solutions to improve in these areas for cluster deployments, and are removing this from the roadmap until we determine the solution and timeline.

If you’re interested in this feature, please share your feedback in the GitHub community so we can track interest and consider it in the future.