Feed behaving differently on manual and auto refresh

problem
1

Hello.

A while ago, one of my feeds started having issues.
The history shows that auto-refreshes get a 403 HTTP error.
However, if I manually request a refresh I get a 200 OK.

The feed URL is Journal du Geek

None of my other feeds have this issue.
Perhaps the client request is different on each method, and that website filter them out.

I use the newsblur website interface exclusively.

Any ideas?

2

Hello.

The problem persists. Is there anything I should do to troubleshoot it?

3

Hello.

The reported issue is still there.
I am forced to manually go in site settings and trigger a refresh every time.
Fetches only succeed when I trigger them that way.
Fetch log looks like this:

[Parse this RSS/XML Feed]
2025-02-01 09:41:16 OK (200)
2025-02-01 09:03:31 HTTP Error (403)
2025-02-01 06:31:28 HTTP Error (403)
2025-02-01 04:03:23 HTTP Error (403)
2025-02-01 01:52:52 HTTP Error (403)
2025-01-31 23:47:15 HTTP Error (403)

There must be a distinction between these two types of requests. to trigger that difference.
Whether it’s the use of IPv4/6, the source IP, a TLS negotiation difference or an HTTP header.

For instance, I have once narrowed down the cause of such a behavior on another website to the use of ALPN. That website is routed by Cloudflare, but other are as well and don’t present that behavior.

As it happens, the domain www.journaldugeek.com is also routed by cloudflare. So, maybe that’s a hint.

Anything I can do to draw the attention of the maintainer?

4

The difference is the server used to fetch the feed. The 403 is hitting NewsBlur’s task servers which fetch feeds in the backup. The 200 is from the app servers. You can make this faster by hitting capital-R while in a feed to insta-fetch stories. But I’m working on a solution to get those stories in the background feed fetchers without you having to reach out to the publisher.

5

Ha, thank you for the hint! It’s very helpful.