GitHub user kruscho added a comment to the discussion: Monitoring Adapters
Hi @tenthe, maybe I was thinking too simply. We already know the number of events per adapter. By having this number + adaptername available through prometheus , we can have a closer look on the behavior of adapters. I completely agree with you regarding the abnormal behavior. It is hard to detect, if zero events are "normal" or a drift. Maybe we can throw some AI on it, to detect normal and anormal manner :-). My idea was more like knowing the time period when an adapters sends nothing (and going a posteriori in a debugging). Something like a calendar, where we can manage planned downtime could be a kind of a solution. But to be honest, i don´t think, this calendar will be maintained really good. Personally, I would be happy to have the same status like pipelines available in prometheus. Maybe next step would be to have eventnumber (as throughput) and adaptername (or element) available. I would let the users/admins decide to define (a)normal metrics GitHub link: https://github.com/apache/streampipes/discussions/3936#discussioncomment-15066240 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
