However, my bet is we will need to make one modification, since we are discarding response bodies. Instead of res.body.length, I imagine we will instead want to look at res.headers['Content-Kength']. (I've confirmed this is set correctly on both raw & boost http requests.
We'll also want to make sure we can get bandwidth from the combination of data_received/duration -- maybe it makes sense to just record this as an additional metric rather than post-hoc.
Acceptance criteria
We can generate a mean / median / 90% for bandwidth for boost vs raw from the saved summary file at the end of the test.
Goals
We want to be able to seperate bandwidth measurement between boost HTTP requests and raw HTTP requests.
The built-in "data received" param doesn't allow us to do this. However, we should be able to do this with a mechanism described here: https://k6.io/docs/examples/track-transmitted-data-per-url/
However, my bet is we will need to make one modification, since we are discarding response bodies. Instead of
res.body.length
, I imagine we will instead want to look atres.headers['Content-Kength']
. (I've confirmed this is set correctly on both raw & boost http requests.We'll also want to make sure we can get bandwidth from the combination of data_received/duration -- maybe it makes sense to just record this as an additional metric rather than post-hoc.
Acceptance criteria
We can generate a mean / median / 90% for bandwidth for boost vs raw from the saved summary file at the end of the test.