Wesley Pettit

@mwarzynski > As for the tests, feel free to check my tests implementation and provide your opinion, but I'm afraid the PR of the tests contains too many changes to...

> I have two Valgrind complaints in the current version. Sorry I am not sure what caused these @mwarzynski... if Valgrind reports no memory leaks then that's the main thing...

@mwarzynski with regards to testing, my main concern is that prior to any release we must have BOTH the implementation fully completed and tested in the real world AND unit...

@mwarzynski Can you please squash this to a single or a few commits as we are getting closer to merge?

@mwarzynski Also I just tried testing it myself and it doesn't work: FB output: ``` $ ./bin/fluent-bit -c ~/configs/aws.conf Fluent Bit v2.0.2 * Copyright (C) 2015-2022 The Fluent Bit Authors...

@Sh4d1 Do you think anyone would ever want to connect to IMDS over a proxy? I think no... that seems silly and unsafe. If you have a proxy configured though,...

@matthewfala I think we should probably remove the check you added... right now with how the proxying works, the user has to set IMDS in the no_proxy settings: https://github.com/fluent/fluent-bit/blob/master/src/flb_upstream.c#L245 @edsiper...

> is there any action pending on my side here ? @edsiper I don't think so. We can submit the PR.

I think the solution here is to refactor this code into 2 major functions: https://github.com/fluent/fluent-bit/blob/master/src/flb_upstream.c#L222 * `flb_upstream_create`: existing functionality that contains proxy settings * `flb_upstream_create_no_proxy`: same thing but without all...

@edsiper IMO, this is a pretty major miss in the `cri` parser.