feat(inputs.system): Add operating-system information

[bilkoua]

Summary

This is the second step of the plan agreed in #18533: add an os value
to the include option of inputs.system that emits a separate
system_os measurement with operating system release and uname
information.

The data is gathered through gopsutil for cross-platform support and
is read with the granular host.PlatformInformation(),
host.KernelVersion() and host.KernelArch() calls instead of
host.Info() to avoid unrelated probes for virtualization, boot time
and process counts.

The result is cached for os_cache_ttl (default 5m) between gathers
since these values are effectively static at runtime; setting
os_cache_ttl = "0s" keeps the cache for the lifetime of the
process, while a tiny positive value such as "1ns" effectively
disables the cache and forces a fresh read on every gather. The TTL
handling matches the convention used by inputs.sqlserver,
inputs.logql, inputs.promql, outputs.iotdb and
outputs.parquet where a zero duration disables the time bound.

Fields emitted by the system_os measurement

Field	Type	Source
os	string	runtime.GOOS
platform	string	/etc/os-release etc.
platform_family	string	/etc/os-release etc.
platform_version	string	/etc/os-release etc.
kernel_version	string	uname -r
kernel_arch	string	uname -m

On platforms where gopsutil cannot provide a particular value (e.g.
parts of FreeBSD/OpenBSD/Solaris) the corresponding field is left
empty; if no field can be gathered the metric is skipped entirely.

Backward compatibility

The default include set is unchanged, so existing deployments keep
emitting only the system measurement and do not see the new
system_os metric until they opt in.

Example output

system_os,host=worker-01 os="linux",platform="ubuntu",platform_family="debian",platform_version="26.04",kernel_version="7.0.0-7-generic",kernel_arch="x86_64" 1748000000000000000

Checklist

• No AI generated code was used in this PR
• AI generated code used in this PR follows the InfluxData Policy on AI-Generated Code Contributions

[bilkoua]

@skartikey @srebhan @Hipska Hi, could you take a look at this PR when you have a moment?

[srebhan]

Thanks @bilkoua for the follow-up PR! Please find my comments below...

[bilkoua]

@srebhan @Hipska Thanks for the review!
Could you take another look?

[srebhan]

Thanks @bilkoua! Some minor comments...

[srebhan]

One more thing I noticed. Sorry for the initial noise. :-(

[srebhan]

Thanks a lot @bilkoua!

[skartikey]

@bilkoua Thanks! A couple of comments to look at.

[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 significantly change the Telegraf binary size (less than 1%)

📦 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

[bilkoua]

@skartikey Thanks for the review!

Addressed both comments. Not 100% sure the test approach is exactly what you had in mind. I went with IgnoreFields for arch/kernel_version (uname-based, host-dependent) and assert them non-empty separately, while the os-release-parsed values are compared literally. Similar pattern is used in inputs.statsd and inputs.sip.
Could you take another look?

[skartikey]

@bilkoua Thanks for the contribution!

[skartikey]

@skartikey Thanks for the review!

Addressed both comments. Not 100% sure the test approach is exactly what you had in mind. I went with IgnoreFields for arch/kernel_version (uname-based, host-dependent) and assert them non-empty separately, while the os-release-parsed values are compared literally. Similar pattern is used in inputs.statsd and inputs.sip. Could you take another look?

That works for me. The os-release-parsed values are now actually verified, which was the main concern, and the separate non-empty checks for arch/kernel_version cover the host-dependent fields. Thanks for the quick turnaround. Approved.