Several of our services in our environment have to change together (API calls between services). Today each service cuts over on its own healthcheck. So during a release, the new version of one service can serve traffic next to the old version of another. If one service fails its healthcheck, the others are already live on the new version and stay there.

What we’d like:

Put services into a deploy group (for example, by service group).

When any member deploys, or the group is deployed through CLI/API (railway redeploy --group ), every member brings up its new deployment next to the old one.

Traffic switches for all members only when every new deployment passes its healthcheck.

If any member fails or times out, none of them switch, and the whole group stays on the previous version.

Roll back the whole group with one action.

This came up already in 2024 in blue/green thread. Railway’s answer then was that services can’t go out or fail together.

Is this planned any time soon?