Ash framework monitoring

[joshuap]

In addition to telemetry, Ash also supports built-in tracing which is similar to telemetry but can share context across processes:

https://hexdocs.pm/ash/monitoring.html

The AppSignal Tracer implementation (created/maintained by the Ash core team) is here, and looks pretty straightforward:

https://github.com/ash-project/ash_appsignal/blob/main/lib/ash_appsignal.ex

They have a suggested inclusion list of trace types here:

https://github.com/ash-project/ash_appsignal/blob/main/documentation/tutorials/getting-started-with-ash-appsignal.md#configuration

I'm not sure what would work best for Insights—telemetry or tracing. (For those who are unfamiliar, Honeybadger Insights uses flat events instead of child spans, with the ability to correlate context across events).

Also, here's some info about error handling in Ash:

https://hexdocs.pm/ash/error-handling.html
https://hexdocs.pm/ash/Ash.Tracer.html#c:set_error/1

[joshuap]

@rabidpraxis I believe you did some work to track request ids across telemetry events in Insights. I wonder if Ash's tracing could work with that?