You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add multipath collector for NVMe-oF subsystem path health
Add a new disabled-by-default collector (--collector.multipath) that
exposes NVMe over Fabrics connectivity metrics by reading
/sys/class/nvme-subsystem/:
- node_multipath_nvme_subsystem_info: subsystem metadata (NQN, model, etc.)
- node_multipath_nvme_subsystem_paths_total: total controller paths
- node_multipath_nvme_subsystem_paths_live: live controller paths
- node_multipath_nvme_path_state: per-controller state (live/dead/etc.)
This fills a monitoring gap for storage connectivity — the existing NVMe
collector reports hardware health but is blind to fabric path failures.
No special permissions required; sysfs is world-readable.
Signed-off-by: Shirly Radco <sradco@redhat.com>
Co-authored-by: AI Assistant <noreply@cursor.com>
The `multipath` collector exposes NVMe-oF (NVMe over Fabrics) subsystem path
346
+
health by reading `/sys/class/nvme-subsystem/`. It provides connectivity-layer
347
+
visibility that the standard `nvme` collector does not cover — specifically,
348
+
per-controller path state and path redundancy counts at the subsystem level.
349
+
350
+
This enables alerting on fabric path failures (e.g. disconnected FC cables or
351
+
failed switches) before they cause total storage loss.
352
+
353
+
No special permissions are required — the sysfs files are world-readable.
354
+
342
355
### Filtering enabled collectors
343
356
344
357
The `node_exporter` will expose all metrics from enabled collectors by default. This is the recommended way to collect metrics to avoid errors when comparing metrics of different families.
0 commit comments