Azure Isolated Function template: why don't we use the DI container ?

[fgheysels]

In the Isolated Function template, we pre-provision a sample Azure Function.
This function uses a logger, but instead of having that logger injected into the function, we're creating it ourselves. Why do we do that ? Why don't we use the DI container for this ?

In other words, we're currently doing this:

ILogger logger = context.GetLogger<OrderFunction>();
logger.LogInformation("C# HTTP trigger 'order' function processed a request");

(source)

Why don't we just inject the ILogger via the constructor ?

[stijnmoreels]

Ah, probably bc this was also done in the regular Microsoft Azure Functions project template.
I'm guessing the difference here is that the logger will be created using the scoped request services instead of the regular ones, but otherwise, I'll have to look at the code.

So, mainly streamlining with Microsoft default here, I guess.

[fgheysels]

I think a DI approach is more 'best practice', don't you think ?

[stijnmoreels]

Yes, I would think so too, but their templates in both in-process and isolated are retrieving the logger instance in a slightly different way. It is still a logger from the DI container, only retrieved in a scoped fashion.

[fgheysels]

Ok. I meant: we should inject it in the ctor, instead of retrieving it like we do now :)

[stijnmoreels]

😄 , depends on what effect that has on the request logging, maybe. If we still have the same telemetry as before. Otherwise, yeah 👍