@michaln Thank you this will solve the uptime monitoring problem. And thank you for pointing out the StorJ official multi-node dashboard.
@beast Thank you this was very interesting reads, to date I was just looking at Kafka and RabbitMQ for such purposes (of course with quite some coding on top). And it looks like a variable option for the development of a large scale notification service.
From my point of view, the StorJ should not be prioritizing very high the multi-node dashboard as the core is more of a priority and data on the network. And I see the document does cover some of the needs however some stay uncovered (like predictions and logs reading, such I do not think should be prioritized at this point from the StorJ team). The priority is why I see it as an opportunity to make such software as a third party, and essentially help the framework grow without loading the core team.
In the end, I am hoping to create a service that will work with the SNs and provide all the knowledge that is now available in this forum and from various scripts and google sheets.
At current my idea for design is to run a container for every node which goal will be to read the logs from the node, and the current UI API (as there is no other at the moment), and send the data to a web service over HTTPS. For UI, for now, I am considering an Android app as the goal of the service is data when you are not at home, so it seems most appropriate, and it opens the door for the notifications. For the local monitor I think GO will be the best fit as the container will become very small, for the web service, for now, I am looking at Java/Kotlin as the BE code and I am still looking for a good DB solution that can store the statistics data in a good and queriable format, there I prefer if I can store the data on the StorJ network and essentially load the network a bit more. I think this should be possible with Hadoop, but I have more to read on the topic.
With the above, I think the features of success rate detection, usage monitoring, offline detection, income prognoses, and failure detection (from the logs) can be made in a secure manner, without exposing ports of a home network. I am also thinking of a set of hints, for common scenarios like 2 nodes on the same network that are both not full for example, but that will be in the future.
Sorry for the long post, and thank you for your responses.