Diagnostic tags for # of tags per quality per dataset
It would be really helpful if we could have # of tags per quality (or even just good/not good) per dataset. That way we could see this in trends and use it for events/alerts like if we had a collector problem somewhere. We have manually done something like this in Excel and played around with pulling these numbers and publishing them through MQTT back to Canary for storage but a "built in" solution would be much better.
Maybe ideally updated each minute but we could maybe get by with something else like per hour or day even
3 replies
-
Yes! And I don't know if I can lump this in here. But DataSet-level metrics would be nice. For instance last write time. I know you get an email if a DataSet goes stale, but customers these days have enterprise monitoring apps that want an API to query.
Problem is, then you'll start getting into tag count issues. That's potentially a lot of tags for Diagnostic to take away from your license. Maybe you could opt into DataSet level Diagnostics. Or.... hear me out. Diagnostic tags don't count against the license limit. Not 100% how you'd prevent people from abusing that, but, it would be super cool. -
- Great idea. For comparison purposes, OSI PI had an additional Interface that you could add to a Gateway Server called the "Performance Monitor Interface" which would automatically create tags that you could turn on and off for historization. This way you could see live values and then trend the ones that you wanted to see long term.
-
FYI for others who may be interested: We are experimenting with a Node-RED flow (running in FlowFuse) that runs every hour (or whatever frequency we tell it) or on demand from a refresh button on the Node-RED dashboard that uses the Canary API to check all the datasets and tags to get current qualities and send that info back to Canary for historization.
This setup allows
- Quick way to see live or max 1 hour old quality counts without using Excel
- Way to trend tag quality counts
- Events/Email Alerts
Again, this is still experimental and there is some potential overhead from this (and I'm also trying to make sure the call time of this is reasonable (seems to be 10 seconds or so normally) and that it doesn't have any negative impacts on the system in general but I wanted to mention this. I'm happy to share the Node-RED flow if you'd like (just DM me).
Node-RED Dashboard:
Axiom Dashboard:
Print this page