fix(inputs.cisco_telemetry_mdt): Handle NXOS DME subtree telemetry format

[yusufshalaby]

Summary

Fixes the alignment of dn tags with their corresponding fields for telemetry paths on NXOS where query-target=subtree.

Checklist

• No AI generated code was used in this PR

Related issues

resolves #15922

[srebhan]

Thanks @yusufshalaby for your fix! I'm not sure I understand why your code-change should fix the issue? In principle your only actual modification is to delete the dn tag or am I getting this wrong?

[yusufshalaby]

Hi @srebhan, thanks for reviewing this fix!

The main modification is that I'm setting the dn value of the tags map of each DME path group prior to adding the fields to the SeriesGrouper. The plugin currently mutates tags with a new dn and then carries that dn to the next path group, which is why I'm also deleting dn from tags after the loop. The new unit test should help clarify, but here's an example in json to make it easier to read. Suppose this is the input:

[
  {
    "descr": "this is eth1/1",
    "dn": "sys/intf/phys-[eth1/1]",
    "id": "eth1/1"
  },
  {
    "descr": "this is eth1/2",
    "dn": "sys/intf/phys-[eth1/2]",
    "id": "eth1/2"
  },
  {
    "descr": "this is eth1/3",
    "dn": "sys/intf/phys-[eth1/3]",
    "id": "eth1/3"
  }
]

Currently the plugin would parse this input as:

sys/intf descr="this is eth1/1"
sys/intf,dn="sys/intf/phys-eth[1/1]" id="eth1/1",descr="this is eth1/2"
sys/intf,dn="sys/intf/phys-eth[1/2]" id="eth1/2",descr="this is eth1/3"
sys/intf,dn="sys/intf/phys-eth[1/3]" id="eth1/3"

My simple fix is to first loop through the fields to find dn and update tags, then loop through the fields once again and add each field to the SeriesGrouper, ensuring they're all grouped together with the same tags:

sys/intf,dn="sys/intf/phys-eth[1/1]" descr="this is eth1/1",id="eth1/1"
sys/intf,dn="sys/intf/phys-eth[1/2]" descr="this is eth1/2",id="eth1/2"
sys/intf,dn="sys/intf/phys-eth[1/3]" descr="this is eth1/3",id="eth1/3"

[srebhan]

Yeah but I don't see where the dn tag is used between setting it in line 629 and deleting it in line 637! That's why I'm asking. :-) Could you please point out where this is used?!?!

If the dn tag is used transiently, which is the case as you set and delete it unconditionally, it should be passed as a parameter instead of using tags for passing function params.

[yusufshalaby]

tags gets passed to parseContentField method on line 635 which passes tags (containing dn) to the grouper.add() on line 679. Is that what you're missing?

[srebhan]

Yes that's what is missing. And you need to delete the tag to avoid it being carried over to the next node which might not define a dn at all. Is this correct? If so, please add a comment to the code stating the above!

[yusufshalaby]

Yes, exactly. Added comments.

[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

[srebhan]

Thanks for your effort and patience @yusufshalaby!