It'd be awesome if there were more dev posts about hooking into the node runtime, about how to layer in observability especially of code you don't control.
I've occasionally peaked at dd-trace-js. The last two years seem to have been a major overhaul of the whole Node observability core, quickly percolating to the whole ecosystem, with diagnostic channels becoming a thing & everyone rebasing their approach stop it. That's a huge part of the picture, I'd wager, for how info gets recorded. https://nodejs.org/api/diagnostics_channel.html
But also, what Node offers for hooks, how we can observe node & various libraries; there's gotta be hooks & techniques up the wazoo to do that. I don't have any clue how application monitoring libraries do it. JS allows a lot of cool monkey patching but man, but folks (to my chagrin) tend to design much tighter surfaces libraries than in the past. (Personally I find strong encapsulation to be an unfortunate hard bound to force on others.)
I've occasionally peaked at dd-trace-js. The last two years seem to have been a major overhaul of the whole Node observability core, quickly percolating to the whole ecosystem, with diagnostic channels becoming a thing & everyone rebasing their approach stop it. That's a huge part of the picture, I'd wager, for how info gets recorded. https://nodejs.org/api/diagnostics_channel.html
But also, what Node offers for hooks, how we can observe node & various libraries; there's gotta be hooks & techniques up the wazoo to do that. I don't have any clue how application monitoring libraries do it. JS allows a lot of cool monkey patching but man, but folks (to my chagrin) tend to design much tighter surfaces libraries than in the past. (Personally I find strong encapsulation to be an unfortunate hard bound to force on others.)