Why is My YouTube Data Source in Pipedream Receiving Multiple Video Upload Events from Unrelated Channels?

This topic was automatically generated from Slack. You can find the original thread here.

Hello, there appears to be an with the Youtube Data Source, I’m not sure if its a bug, but noting it here in case it is:

One of my YouTube Data Sources [I have two of the same kind: the difference is the scheduling of when they run, due to upload frequency and importance between the two channel sources] randomly received video upload events multiple times from unrelated channels the morning of 19th December at 4am GMT. The source went completely crazy (see image below for the amount of repeated events were received) because of it – the workflow assumed that there would not be an issue of this kind, so the upload was passed straight through to the processing step using Node.js – and I had to turn it off.

This was to prevent a repeat of the problem, whereby the destination – a Discord Webhook posting a message to let people know that a new video was uploaded on a specific channel – was flooded by all the spammy uploads from a completely irrelevant channel that I did not tell Pipedream to subscribe to. It led people to believe the webhook was hacked (it wasn’t) and required us to remove the spam and then hide the channel they were posted to. Considering the same videos were appearing more than one all at the same time, this would lead me to believe the issue was on Pipedream’s side, and not an issue with my configuration of the source.

We never used to have this kind of problem with this flow on Pipedream, but after switching to the new system – due primarily to the prompts in the website and the fact that the legacy workflows used up most of the free credits – we are now getting a lot more issues than I’d like. One of the issues I previously described here was found to be potentially code related which I’ve manually resolved and removed from this post. Thanks

Hi, can you please open a support ticket for this with the link or deployed component ID of the source? Or DM it to me

Hi unfortunately, I’ve since deleted and recreated these sources and wrote code in Node.js that is now designed to ignore any kind of issues from the YouTube Data source from reaching the destination platform. If it does recur again, I will raise a support ticket with the details you’ve advised. Apologies :frowning:

Posted thread to Discourse: Why is my Pipedream YouTube Data Source reporting random video uploads from unrelated channels?

while this particular bug has not reared its head again since, I am now getting a new error where, for no particular or apparent reason, the source starts spamming a JavaScript error in my source logs: this eventually causes Pipedream to disable my source because it has a 100% error rate, which isn’t even caused by me. Please review this.

The data source ID: dc_wDuZVwN

When I re-created this data source prior to Christmas, the errors stopped: however, I do not intend to repeatedly delete and recreate this data source every time the source just starts spamming TypeErrors by itself. (The error being sent to the source logs is "TypeError: Cannot read properties of undefined (reading ‘snippet’). There are no details to suggest where this error is coming from, but since its in the logs of the source, not my workflow, it would suggest the problem is with the source. This issue does NOT impact the other YouTube Data source on my profile. Thanks.

if you need me to raise a support ticket through a specific channel let me know as well, thanks.

Sorry for the delay to respond here - I’m prioritizing this issue for the team to fix. Thanks for reporting it

You can keep track of it here: [BUG] Youtube Data API source throwing errors · Issue #19736 · PipedreamHQ/pipedream

oops, I just checked in here today after just submitting a support ticket which I think posted to the github repo lol. You can delete this ticket as its obviously no longer needed → YouTube Data · Issue #19802 · PipedreamHQ/pipedream · GitHub

I can see the fix is live from your link, so will test it soon!

or at least, I think it is planned for fix, so I think I just need to wait a day or two for it to propagate out to my source?

After updating and redeploying the source, the changes will be applied - this is already published