There's no question that it's a huge step up in complexity to wire together such tightly-linked front and backend components, but it is done for things like GIS, where you want data overlays.
I think it's just a different mindset; GIS libs like Leaflet kind of assume they're the centerpiece of the app and can dictate a bunch of structure around how things are going to work, whereas charting libs benefit a lot more from "just add me to your webpack bundle and call one function with an array and a div ID, I promise not to cause a bunch of integration pain!"
Last time I tried to use it for dashboarding, I found Kibana did extremely aggressive down-sampling to the point that it was averaging out the actual extremes in the data that I needed to see.
I think it's just a different mindset; GIS libs like Leaflet kind of assume they're the centerpiece of the app and can dictate a bunch of structure around how things are going to work, whereas charting libs benefit a lot more from "just add me to your webpack bundle and call one function with an array and a div ID, I promise not to cause a bunch of integration pain!"
Last time I tried to use it for dashboarding, I found Kibana did extremely aggressive down-sampling to the point that it was averaging out the actual extremes in the data that I needed to see.