fix(outputs/graylog): fix failing test due to port already in use

[sspaink]

Direct link: https://app.circleci.com/pipelines/github/influxdata/telegraf/7370/workflows/dd5cc7c4-9b38-4f91-9f6d-86040c4fca5a/jobs/127224/tests

The error message from the CI:

graylog_test.go:167: 
     Error Trace:	graylog_test.go:167
      	            		asm_amd64.s:1581
     Error:      	Received unexpected error:
      	           listen udp 127.0.0.1:12201: bind: address already in use
     Test:       	TestWriteUDP/UDP

Description:

The output plugin graylog tests are failing due to trying to bind to a port already in use, because the previous test didn't fully unbind yet. I found this package: https://github.com/libp2p/go-reuseport that makes it easy to call a syscall SO_REUSEPORT so that the bind can be reused even if something else was already bound to it. Originally I was hoping to just use bytes.Buffer instead to avoid having to start a UDP and TCP server in a test, but it doesn't cover the connection logic. This approach introduces the fewest changes, so decided this would be better.

[telegraf-tiger]

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

📦 Looks like new artifacts were built from this PR.

Expand this list to get them here! 🐯

Artifact URLs

DEB	RPM	TAR GZ	ZIP
amd64.deb	aarch64.rpm	darwin_amd64.tar.gz	windows_amd64.zip
arm64.deb	armel.rpm	freebsd_amd64.tar.gz	windows_i386.zip
armel.deb	armv6hl.rpm	freebsd_armv7.tar.gz
armhf.deb	i386.rpm	freebsd_i386.tar.gz
i386.deb	ppc64le.rpm	linux_amd64.tar.gz
mips.deb	s390x.rpm	linux_arm64.tar.gz
mipsel.deb	x86_64.rpm	linux_armel.tar.gz
ppc64el.deb		linux_armhf.tar.gz
s390x.deb		linux_i386.tar.gz
		linux_mips.tar.gz
		linux_mipsel.tar.gz
		linux_ppc64le.tar.gz
		linux_s390x.tar.gz
		static_linux_amd64.tar.gz

[powersj]

Is it possible to change the port each test uses? That way each test has a unique port number.

[sspaink]

@powersj probably, but I think you do run the risk of another test using the same port number, the tests can run in parallel and in random order so you might end up with a sneaky "already bound" issue. I guess you could keep trying a new port until you find one?

edit: Using ":0" could work as well: https://stackoverflow.com/questions/43424787/how-to-use-next-available-port-in-http-listenandserve but I guess, it could pick a free port and another test could pick the same port. Not sure how to prevent collisions. Maybe these tests should be marked as integration tests.