Our production application experienced widespread connectivity issues for approximately 32 minutes on January 14, 2026, from 18:05 to 18:37 UTC. Many/most of our users reported slow loading screens and "Something went wrong" error messages. Investigation reveals evidence of network connectivity failures affecting inbound connections from Cloudflare to Railway.

• 18:05-18:37 UTC: 39 "BadRequestError: request aborted" errors in application logs

• 18:05-18:37 UTC: Multiple Cloudflare 522, 524, and 525 errors reported to users (captured in Sentry)

• 18:37 UTC: Errors ceased, normal operation resumed

Cloudflare Error Codes (from Sentry frontend monitoring):

• HTTP 522 (Connection timed out): 23 events affecting 15 users

  • "Cloudflare could not negotiate a TCP connection with the origin server"

• HTTP 524 (A timeout occurred): 9 events affecting 4 users

  • "Cloudflare successfully connected but the origin server didn't respond in time"

• HTTP 525 (SSL Handshake Failed): 19 events affecting 6 users

  • "Cloudflare could not negotiate a SSL/TLS handshake with the origin server"

We're a B2B SaaS for enterprise customers with 10-15 users active at any one time so the figures quoted are statistically significant.

Errors in Railway application logs:

39 instances of the following error during the incident window:

BadRequestError: request aborted
    at IncomingMessage.onAborted (/app/node_modules/.pnpm/raw-body@3.0.1/node_modules/raw-body/index.js:245:10)
    at /app/node_modules/.pnpm/@opentelemetry+context-async-hooks@1.30.1_@opentelemetry+api@1.9.0/node_modules/@opentelemetry/context-async-hooks/build/src/AbstractAsyncHooksContextManager.js:50:55
    at AsyncLocalStorage.run (node:internal/async_local_storage/async_hooks:91:14)
    at SentryContextManager.with (/app/node_modules/.pnpm/@opentelemetry+context-async-hooks@1.30.1_@opentelemetry+api@1.9.0/node_modules/@opentelemetry/context-async-hooks/build/src/AsyncLocalStorageContextManager.js:33:40)
    at SentryContextManager.with (file:///app/node_modules/.pnpm/@sentry+opentelemetry@9.47.1_@opentelemetry+api@1.9.0_@opentelemetry+context-async-hook_31f1e432282b85e8bf960b6aef555cb2/node_modules/@sentry/opentelemetry/build/esm/index.js:1430:24)
    at IncomingMessage.contextWrapper (/app/node_modules/.pnpm/@opentelemetry+context-async-hooks@1.30.1_@opentelemetry+api@1.9.0/node_modules/@opentelemetry/context-async-hooks/build/src/AbstractAsyncHooksContextManager.js:50:32)
    at IncomingMessage.emit (node:events:518:28)
    at IncomingMessage._destroy (node:_http_incoming:221:10)
    at _destroy (node:internal/streams/destroy:122:10)
    at IncomingMessage.destroy (node:internal/streams/destroy:84:5)

These errors have no corresponding request logs in our application middleware. Our Express server logs every incoming request with a unique ID, but these 39 aborted requests never reached our logging middleware. This indicates connections were terminated at the network/proxy level before reaching our Express application.

Infrastructure Setup

• Hosting: Railway

• CDN/Proxy: Cloudflare (proxying all traffic to Railway)

• Application: Next.js (Pages Router) with custom Express server

• Database: MongoDB/Mongoose

Could you please investigate whether there was a network connectivity issue affecting our Railway deployment during this timeframe? We do also periodically see similar slowness and errors while hosting on Railway so it's not completely a one-off but this makes a useful "case study" to investigate the issue