Starting approximately 2 days ago (around March 6, 2026), all my Next.js services are returning "Application failed to respond" (502 Bad

Gateway) on the first request after the app has been sleeping. Refreshing the page immediately after resolves the issue — the app loads

fine once awake.

Key details:

- Framework: Next.js 16 (App Router, standalone output mode)

- Affected services: 3 Next.js apps (home, client, admin) across both staging and production environments (6 services total)

- All services have sleep enabled

- No deployments to production in over a month — this rules out a code change on my end

- Staging has had recent deploys, but the same issue also appears on production with no code changes

- Behavior: First request after sleep → 502. Manual refresh → works fine. This is consistent and reproducible across all 6 services.

- Started: ~2 days ago. Before that, sleeping apps woke up without any 502 errors.

- Request ID (example): X5dEd7jNS3uwNtjNw9P4nw

What I suspect:

Next.js standalone server takes a few seconds to cold-start. It seems like Railway's proxy is now returning a 502 before the Next.js

server finishes booting — rather than waiting for it as it did previously. This may be related to a change in proxy timeout or sleep

wake-up handling, or possibly a change in how Railway handles Next.js apps specifically.

What I've ruled out:

- No code or config changes on production

- All 6 services (3 apps x 2 environments) exhibit the same behavior simultaneously

- Apps work perfectly once awake — no errors in deploy logs

- Not app-specific — affects home (marketing), client, and admin apps equally

Could the Railway team check if there was a recent infrastructure change affecting sleep wake-up timeouts, proxy behavior, or

Next.js-specific handling? Thanks!