[BUG] Disk integration `system.disk.read_time` `write_time` `system.io.await` missing mount name

[CrashLaker]

Hi all,
idk if this is a bug or a FR. if not. apologies

Agent Environment
version 7.52.1

Describe what happened:
I've a dashboard where I monitor my disk pools.

I have many mounts like

the way i setup the panels are like this

so I use the mount name /mnt/data-YRLB for disk size
and /dev/sdd1 for io latency.

and every time the host reboots the device sd* scrambles to another mount. since they're bounded by their UUIDs everything works fine but the dashboard gets out of sync.

Describe what you expected:
I'd expect to use the mount name in metrics like system.disk.read_time, system.disk.write_time, system.io.await
like in system.disk.used

Steps to reproduce the issue:
start agent and enable disk monitoring

Additional environment details (Operating System, Cloud provider, etc):
centos 8
on bare metal mac mini

thanks

regards,c.

[CrashLaker]

well not ideal but i've edited /opt/datadog-agent/embedded/lib/python3.11/site-packages/datadog_checks/disk/disk.py

    def collect_latency_metrics(self):
        output = os.popen("mount").read().strip()
        # /dev/sdi1 on /mnt/data-3RJT type ext4 (rw,relatime)
        mapmount = {}
        for line in output.split("\n"):
            try:
                l = line.split(" ")
                if len(l) >= 3 and l[1] == "on":
                    mapmount[l[0]] = l[2]
            except:
                pass

...

                dev_disk_name = f"/dev/{disk_name}"
                if dev_disk_name in mapmount:
                    metric_tags.append('device:{}'.format(mapmount[dev_disk_name]))
                else:
                    metric_tags.append('device:{}'.format(disk_name))

maybe a feature request.

regards,c

[FlorentClarret]

Hi @CrashLaker, thanks for reporting this! I'm moving this issue to the integrations-core repo where the integration lives so the team in charge of it can have a look.