a month ago
Project: gomo-web (30fb366a-9021-45e1-9827-50f3e040d2c4)
Environment: production (2b38aff3-0257-4dbc-bf6e-7d23ee1b3cf2)
Service: admin (5b127af9-35b6-4e2a-9c4f-6460487a1ad0)
Region: EU West (europe-west4-drams3a)
Symptom
The admin service is unreachable over private networking. The web+api service proxies requests to http://admin.railway.internal:3000, but all requests return 502 because the hostname won't resolve.
Diagnosis
railway private-network status --service admin --json output:
- endpoint: admin.railway.internal (ID: e69c46ef-3157-4795-ba93-380e9b9a3a25)
- syncStatus: "ACTIVE"
- privateIps: [] (EMPTY — no IP allocated)
For comparison, web+api endpoint (restu-web.railway.internal) on the same network correctly has privateIps: ["fd12:58b3:c608:1:1000:4:54ee:a56a"]
Timeline
- Worked until redeploy on 2026-08-25 01:12 UTC (deployment f7fc641b)
- Current: ee582803-b588-4e67-979a-2e545b789de7 (2026-08-25 13:02 UTC)
Attempts (all failed)
- railway service restart --service admin
- railway redeploy --service admin
- railway redeploy --service admin --from-commit ee582803
Endpoint appears stuck: marked ACTIVE but no IP allocated. This looks like an infrastructure-side issue requiring stacker-level operation. Can you re-allocate the private IP?
Pinned Solution
a month ago
If the admin service's ip is being resolved to an IPv6 address, you need to make sure the admin service is binding to IPv6 (::), and not only IPv4 (0.0.0.0).
1 Replies
Status changed to Awaiting Railway Response Railway • about 1 month ago
a month ago
This thread has been opened as a bounty so the community can help solve it.
Status changed to Open Railway • about 1 month ago
a month ago
If the admin service's ip is being resolved to an IPv6 address, you need to make sure the admin service is binding to IPv6 (::), and not only IPv4 (0.0.0.0).
Status changed to Solved 0x5b62656e5d • about 1 month ago