Historical HTTP logs fail while application logs work
diogoup26
HOBBYOP

a month ago

Hello Railway Support,

I cannot retrieve historical HTTP logs for my production service, although recent HTTP logs and historical application logs remain accessible.

Service: mcp-x402

Environment: production

Deployment ID: c06c74da-fa37-4706-894b-5d7cbb47a62e

Operating system: Windows 11 / PowerShell

Railway CLI: 5.43.2

This command fails:

railway.cmd logs c06c74da-fa37-4706-894b-5d7cbb47a62e --http --service mcp-x402 --environment production --since 2026-08-24T15:44:11Z --until 2026-08-24T16:14:11Z --json

Result:

Problem processing request

The same error occurs:

  • With 500, 1,000 and 3,000 lines.
  • With and without the deployment ID.
  • With --since and --until.
  • With --until only.
  • With a narrow 30-minute interval.
  • After upgrading Railway CLI from 5.41.2 to 5.43.2.

Control tests:

  • Retrieving the latest 500 HTTP logs works.
  • Historical application logs for the same service work, including an export of 1,000 lines.
  • The deployment is successful and the service is operating normally.
  • The requested logs are approximately two days old and should be within the Hobby plan’s documented retention period.

This appears isolated to Railway’s historical HTTP-log query. I receive no HTTP 401 or 403 error, only “Problem processing request”.

I need these HTTP logs to correlate application events with Railway edge information, including request IDs, source IPs, user agents, regions, latency and requests that may have failed before reaching the application.

Could you please:

  1. Confirm whether this is a known CLI or HTTP-log backend issue?
  2. Confirm whether the historical HTTP logs still exist?
  3. Provide a workaround or alternative export method?
  4. Help me retrieve the logs for the affected period?

Thank you.

Captura de ecrã 2026-08-24 192113.png

Attachments

$10 Bounty

3 Replies

Railway
BOT

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


aruvine
FREE

a month ago

edit: I had a reply earlier but i removed it now sorry about that


diogoup26
HOBBYOP

a month ago

Thank you. I completed the isolation tests and reproduced the issue on the latest Railway CLI version, 5.43.3.

The following HTTP-log queries work:

  • --since 30m
  • --since 2h
  • --since 6h
  • --since 1d

All were tested with --lines 50, without --json.

The following queries fail with Problem processing request:

  • --since 2h --until 1h
  • --until 1h

The failure therefore occurs as soon as --until is included, even when using relative timestamps, no JSON output and no explicit deployment ID.

Captura de ecrã 2026-08-25 042922.png

Previously, an absolute ISO 8601 interval using --since and --until also failed, both with and without the explicit deployment ID. Historical application-log queries using time ranges work.

I upgraded the CLI from 5.43.2 to 5.43.3 and repeated the --until 1h test; the result remains Problem processing request.

This rules out PowerShell parsing, ISO timestamp formatting, JSON output and an outdated CLI version. It appears to isolate the problem to Railway’s HTTP-log backend/API handling of the until/end-time parameter.

Could this please be reproduced and escalated as an HTTP-log backend issue? I would also appreciate a supported workaround for exporting a bounded historical HTTP-log interval.

Captura de ecrã 2026-08-25 042741.png

Captura de ecrã 2026-08-25 042820.png


aruvine
FREE

a month ago

edit: I had a reply earlier but i removed it now sorry about that


Welcome!

Sign in to your Railway account to join the conversation.

Loading...