-
Notifications
You must be signed in to change notification settings - Fork 14
Description
At the moment, when running a PARAMETRIC
test locally, my console log output becomes completely filled with noise (which seems to include a lot of things, for example, the std. output of the weblog, which for Java, means the logs of the tracers is included).
This creates a bad user experience, because I have to do a ton of scrolling to figure out if my test even failed and if so why.
I think it would be better if I could immediately see the failure, and if there is a failure, it could then direct me to the files in the logs_*
folders I could look at to see the relevant things (request/response bodies, tracer logs, etc).
I realize that this may be suboptimal for the CI use case, where the user would be forced to download the logs_*
folder as an artifact to debug the issue. I can offer a few possible solutions for that problem:
- Have different behavior between running locally and CI, where in CI it would still flood the std. output
- Provide instructions on how to download the
logs_
folder for failed tests from CI - Change the behavior in both CI and locally so that all the logs will be written to Datadog by default, and that would become the default way of investigating failures.