Eugene Krapivin

Well I'd suggest: * [Metrics](https://learn.microsoft.com/en-us/dotnet/core/diagnostics/metrics-instrumentation) - wrapping IO operations in timers (histograms of `TimeSpan` in my code examples) * adding exception/error counters with appropriate tagging * [Tracing](https://learn.microsoft.com/en-us/dotnet/core/diagnostics/distributed-tracing#getting-started-for-net-library-developers) - wrapping IO...

> Right, but how should MailKit provide this info to the app? it should not concern the library. As long as you use `System.Diagnostics.Metrics.Meter` you and document properly the name...

@jstedfast sorry for the late reply. I'd add: * Metrics: * histograms to measure any operation time in the library (like SMTP connect, send, no op, disconnect, etc'). * counters...

instead of using `faults` use a tag/label `status=failed` it will be easier to dissect the metrics using tags. if you go back to my earlier example, you'd see how I...

> That still doesn't answer the most important question which is should these instruments (Counters/Histograms) be per-instance or global? I'm sorry missed that question. I think you should share the...

> I saw that in the docs. I'm not sure if it's worth reporting the network.peer.address, but other than the socket.error code, I'm following the HttpClient pattern. I think starting...

get a beta nuget I could run it through its paces in a demo app I'm preparing for a lecture about observability :) I hope I'm not too lazy I'd...

Wow AWESOME :)

``` ``` looks like it still doesn't really target the latest asp.net?

``` ``` looks like it still doesn't really target the latest asp.net?