Refreshing Feedly's Cache of RSS Feed

148 views
Skip to first unread message

Marcus S

unread,
Aug 5, 2021, 10:39:27 PM8/5/21
to Feedly Cloud
messaged the Feedly general Slack channel on this topic, but am thinking it's more relevant to ask in this Developer Forum. Cross-posting below:

I'm working on creating a RSS feed that Feedly can read and was wondering if there's a way to mark the RSS to invalidate Feedly's cached version of it and prompt Feedly to refetch either certain RSS elements or the entire feed anew. The problem I'm encountering is images aren't loading because their CDN links have expired (and I don't have much control over their expiration time). If I could prompt Feedly to refetch the feed from scratch (rather than what I think is default behavior of polling but not necessarily replacing the RSS with a fresh feed), I believe the CDN links would be regenerated and images would load properly.

I'm at the same time trying to extend the expiration time of these CDN links but it would still help to have a better understanding of how the Feedly cache works.

But regardless, if there's a way to prompt a "fetch from scratch" or "fetch from scratch after 3 days" through some RSS markup, I think that would help ensure photos load properly in my RSS feed.

Am also now aware there was this Q&A https://groups.google.com/g/feedly-cloud/c/3evZeYOnS2I/m/kz1jyORPAwAJ in the developer forum but am wondering if there's any other way to prompt a cache refresh through the XML

David Chatenay

unread,
Aug 12, 2021, 5:40:01 AM8/12/21
to Marcus S, Feedly Cloud
  Hi Marcus,
  Thanks for reaching out.
  RSS has limited support for updates (and none for deleting). It was designed for a "latest news" model, where a limited number (20-50 typically) of recently published articles is provided, not the full timeline.
  In the "poll" model, RSS pollers are expected to fetch RSS feeds regularly, use GUIDs to identify new entries not already fetched, and update existing entries (based on that same GUID) if the content has been updated. If an old entry needs to be updated it would have to be "re-published" in the list of new entries, using the same GUID to avoid creating duplicates. 
  The "push" model (PubSubHubBub) behaves much the same way, except only new and updated articles are pushed to registered servers, which is a lot more efficient and fast.

  In theory, you could provide the entire timeline in your RSS feed. In practice, this wouldn't scale if you have hundreds of articles, which would produce a gigantic XML file. Feedly pollers typically truncate very large feeds and do not process the full dataset.

  So back to your question: if you re-publish articles, the content in Feedly should automatically be updated, assuming that the GUIDs didn't change, the image links were updated, and the Feedly server detects that the content changed. You will need to work in batches (40-50 articles at a time), and wait for the content to be polled/pushed before you move to the next batch.

  In any case, I highly recommend testing the whole process with a single article first, to make sure it gets updated and there is no duplicate article created.

  I hope this helps.

-- 
David Chatenay
Platform Engineer
Feedly


--
You received this message because you are subscribed to the Google Groups "Feedly Cloud" group.
To unsubscribe from this group and stop receiving emails from it, send an email to feedly-cloud...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/feedly-cloud/258701aa-3eb5-462e-ad98-332fa742f03cn%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages