Serilog VS Rebus

To learn more about Serilog, you can visit the documentation on their GitHub project. To learn more about Seq, you can visit the documentation on their website.

These projects use dependency injection for connections and have no logging although they are ready to log to the console via SeriLog setup in Program.cs.

See here: https://github.com/serilog/serilog/wiki/Configuration-Basics e.g.

Captured in public static void OnActivated() and note Log is SeriLog.

"Serilog": { "Using": [ "Serilog.Sinks.File" ], "Enrich": [ "WithHttpRequestId", "WithExceptionDetails", "WithClientIp" ], "MinimumLevel": { "Default": "Information" }, "WriteTo": [ { "Name": "File", "Args": { "path": "../../../../../Logs/MyApp/MyApp-.log", "rollingInterval": "Day", "retainedFileCountLimit": "7", "outputTemplate": "[{Timestamp:yyyy-MM-dd HH:mm:ss.fff zzz} {CorrelationId} {Level:u3}] {Username} {Message:lj}{NewLine}{Exception} {Properties:lj}{NewLine}" } } ] } 👆 I got the enricher names from serilog's wiki. The idea was to log the HTTPRequestId, ClientIp, and automagically log the details of every exception thrown (without having to do _logger.LogError(...)). However I don't see those details in the logs. In Program.cs: ``` var logger = new LoggerConfiguration() .ReadFrom.Configuration(builder.Configuration) .Enrich.FromLogContext() .CreateLogger();

You could also write a custom destructuring policy, assuming it's a flurl-specific public exception type and you only need certain properties of it. You'd implement IDestructuringPolicy and then add it to the configuration:

Serilog for logging

I've cloned https://github.com/serilog/serilog and I'm not sure what's going on with the multitargetting in the tests? Does this (below) mean, if the dev's OS is windows, then target framework 462 or 471? What should I do on my laptop, which only has the net6 and fw48 sdk installed?

I can recommend Rebus: https://github.com/rebus-org/Rebus – it’s the one I know best, because I made most of it 🤠 and it even has a Pro option, if you feel like upgrading in that direction too some time in the future.

The whole trick is that that's the headache of the implementor of the abstraction 😅 (e.g. me, if you're using Rebus).

Rebus (https://github.com/rebus-org/Rebus)

For multiprocess application (multiple services, containerized and also natives) I've used ReBus (https://github.com/rebus-org/Rebus) which was is API-compatible (for some version) with NServiceBus. We first used NServiceBus with RabbitMQ (Sagas stored in SqlServer) and then after migration to ReBus also messages and queues were in SqlServer.

Then we add the Rebus and Rebus.ServiceProvider NuGet packages to each project, giving us something like this: