The pipeline was stable for the last 9 hours, no delays in extraction, everything is working as before.
We will get back to it on Monday to do more research and provide a long lived solution.

For the last several hours, we have had spikes in the queues for article extraction. None of the articles is lost, but we are still struggling to identify the reasons for those spikes. I am afraid that we will continue to see it over the weekend. Our team is constantly working on identifying the issue and providing the fix. But it seems like a much bigger problem than usual.

The issue is resolved. It was not the CloudFLare issue, it was our internal issue. We will review more in details the current issue, create alarms and make sure it does not happen again.

https://www.cloudflarestatus.com/
We fear that it is because of the CloudFlare maintenance that our News API is ddependent on

We are currently experiencing problems with our News API. Most of the calls are timing out.

Investigating the issue, nothing seems off from the server side. We think it might be networking.

Fully resolved

Several hours ago we had an issue with ourr NLP service. Usually we find an article, ingest it in our database and in parallel calculating NLP fields. But due to a code bug from our side, this process was delayed. Currently everything is fixed and we have a queue of messages that we need to process. Over the next 2 hours, we will empty the queue and everything comes back to normal. Keep you posted.