chore(tests): Enable verbose output for tests

[zak-pawel]

Summary

To see the test results for the logger package.

Checklist

• No AI generated code was used in this PR

[telegraf-tiger]

Download PR build artifacts for linux_amd64.tar.gz, darwin_arm64.tar.gz, and windows_amd64.zip.
Downloads for additional architectures and packages are available below.

👍 This pull request doesn't change the Telegraf binary size

📦 Click here to get additional PR build artifacts

Artifact URLs

DEB	RPM	TAR GZ	ZIP
amd64.deb	aarch64.rpm	darwin_amd64.tar.gz	windows_amd64.zip
arm64.deb	armel.rpm	darwin_arm64.tar.gz	windows_arm64.zip
armel.deb	armv6hl.rpm	freebsd_amd64.tar.gz	windows_i386.zip
armhf.deb	i386.rpm	freebsd_armv7.tar.gz
i386.deb	ppc64le.rpm	freebsd_i386.tar.gz
mips.deb	riscv64.rpm	linux_amd64.tar.gz
mipsel.deb	s390x.rpm	linux_arm64.tar.gz
ppc64el.deb	x86_64.rpm	linux_armel.tar.gz
riscv64.deb		linux_armhf.tar.gz
s390x.deb		linux_i386.tar.gz
		linux_mips.tar.gz
		linux_mipsel.tar.gz
		linux_ppc64le.tar.gz
		linux_riscv64.tar.gz
		linux_s390x.tar.gz

[zak-pawel]

It seems to me that something is wrong with the logger package and/or the gotestsum tool.

Here is the test of the logger package on Linux using go test:

As we can see, 21 tests intended for Linux were conducted across three different files (logger_test.go, structured_logger_test.go, and text_logger_test.go).
There are a total of 21 test methods, all of which passed, so everything looks fine.

Now, testing the same package (also on Linux) using gotestsum:

We see that only 4 tests were conducted - both from logger_test.go and the first 2 from structured_logger_test.go.
All 4 tests passed, and the test itself ends "successfully."

But what about the remaining 17 tests?

(I ran the tests on a fresh master branch without any modifications.)

[zak-pawel]

What about CI?

Let's take the latest nightly run on the master branch as an example: https://app.circleci.com/pipelines/github/influxdata/telegraf/24207/workflows/09a89f10-312d-4e77-8681-7123de0b805a

Lets analyze the test logs from all platforms (test-go-windows, test-go-mac, test-go-linux-386, test-go-linux).

Only in test-go-windows we see any information about tests in the logger package, and that’s just that 2 tests were skipped because: "Skipping integration test in short mode."
(Interestingly, both of these tests are from event_logger_test.go, which is Windows-specific. Since we never run integration tests on Windows, these tests are never actually executed.)

For the remaining platforms, there is no trace that the logger package even exists or has any tests.

The same issue with the logger package not being tested occurs in non-nightly tests on the master branch (https://app.circleci.com/pipelines/github/influxdata/telegraf/24172/workflows/fee4ffb2-da00-4254-9fd4-05739e8f4ff3) and in tests triggered by PRs (https://app.circleci.com/pipelines/github/influxdata/telegraf?branch=pull%2F16423).

(In both cases, the test-integration job is also executed, but there is no information about the logger package being tested there either.)

[zak-pawel]

In this PR, I'm making only one change: adding the --format standard-verbose flag to gotestsum.
https://github.com/gotestyourself/gotestsum?tab=readme-ov-file#output-format

Let's look at the CI https://app.circleci.com/pipelines/github/influxdata/telegraf/24201/workflows/acc1ff07-45c3-4ddd-b3c1-b7d28569c13e:

Finally, some information about the logger package appears. For all platforms, only 4 tests are executed in the logger package (the same ones I mentioned earlier).

When I run tests for this package in GoLand, I get the same situation - 4 tests pass, but the job is automatically "terminated."

[zak-pawel]

@srebhan What do you think?

[srebhan]

@zak-pawel TBH I'm not so much in favor of adding more output for cases where the tests pass. However, I'm really concerned about the tests not being run... Or are they? If I do gotestsum --format standard-verbos -- -v -race -short I see all 21 tests are running, while without the -v I only see the 4 ones you mention...

So what is the purpose of adding the flag? Testing? Or do you want to also do it "in production"?

[zak-pawel]

@srebhan

So what is the purpose of adding the flag? Testing?

Testing, just want to understand what is going on with this package.

If I do gotestsum --format standard-verbos -- -v -race -short I see all 21 tests are running, while without the -v I only see the 4 ones you mention...