Railway is saying that the issue should be fully resolved.

While services have returned online, and database connectivity has returned, Railway as of 1:30 am ET is not listing the incident as resolved. Services returned online around 10:30 pm ET.

Railway has successfully restored portions of its service; database connections remain inoperable. As a non-enterprise deployment, our deployments are being deprioritized to keep the build infrastructure stable. There is still no ETA for full recovery.

Railway customer reports suggest that the service outage is primarily affecting Bare Metal-based deployments. Some MCPs were transitioned to Railway's metal infrastructure, but not all. This may explain why Hal was still able to call out during the outage, since that uses Ventus, which we believe is not on Railway's Bare Metal.

The Railway deployment platform is currently experiencing a significant outage. This has caused our three internal MCP servers (Nimbus, Cirrus, and Altus) to go offline, as they are hosted on Railway.

All customer-facing services remain fully operational:
• Cirrusly Weather Front Page ✅
• Cirrusly Weather Shopping Cart ✅
• Cirrusly Weather REST API ✅
• AltWire Front Page ✅
• Matomo ✅

This is not a service outage. We are monitoring the Railway incident and will restore the MCP servers as soon as Railway comes back online. No customer data or functionality is affected.