I’m trying to test gzip compression performance improvements on HTTP(S) traffic with k6. However, some observations does not make sense when using --http-debug option for k6. It looks like (based on --http-debug output) k6 is implicitly adding field “Accept-Encoding: gzip” into http header. However, if I add that field explicitly the order of HTTP header fields are changed, but the header is otherwise the same, yet the test results differ. So, I’m very confused and would like to confirm the result by Fiddler, but I’m not able to capture k6 traffic by Fiddler.
Any idea on how to verify the true http header by Fiddler or some other 3rd party tool because I simply can’t trust the output when using k6 --http-debug. I need a way to verify the output.
I have tried capturing k6 traffic with Fiddler in 1) Win10 command line, using npm scripts to run k6 and 2) Linux (Ubuntu) subsystem in Windows 10, k6 installed on Ubuntu. But capturing with Fiddler fails in both options. I have tried to to configure HTTPS proxy in Windows command line and Ubuntu, but it doesn’t help.