]> git-server-git.apps.pok.os.sepia.ceph.com Git - ceph.git/commit
monitoring: fix hardware Grafana dashboard and health metrics
authorAfreen Misbah <afreen@ibm.com>
Thu, 25 Jun 2026 08:18:19 +0000 (13:48 +0530)
committerAfreen Misbah <afreen@ibm.com>
Thu, 9 Jul 2026 20:11:30 +0000 (01:41 +0530)
commit281a5f9c39f1cfbcfa0d7de41e8bb219b56f4440
treecb191b78b22f2f11da63e56722a14dcecc4264b4
parent285ccc1d5991673d08d2a712f64e653400fff4f1
monitoring: fix hardware Grafana dashboard and health metrics

- Fix fan repeating panels: set multi=true on fan_speeds template
  variable so Grafana generates one panel per fan instead of one
- Remove TACH-only regex filter on fan_speeds template and AVG
  Cooling query so all system fans are visible regardless of naming
- Replace duplicate Power Control panel with Network health panel
- Fix NVMe drive count to use storage_capacity_bytes{protocol=NVMe}
  instead of counting temperature sensors (inaccurate proxy)
- Normalize all hostname filters to regex match (=~) for consistency
- Register hardware.libsonnet in dashboards.libsonnet so the
  dashboard JSON is generated during ceph-mixin builds
- Add temperatures category to prometheus health metrics loop

Signed-off-by: Afreen Misbah <afreen@ibm.com>
Assisted-by: Claude
Signed-off-by: Afreen Misbah <afreen@ibm.com>
(cherry picked from commit 04207ca6dcfc86c51845912848d9cbf1a7cb87d0)
monitoring/ceph-mixin/dashboards.libsonnet
monitoring/ceph-mixin/dashboards/hardware.libsonnet
monitoring/ceph-mixin/dashboards_out/hardware.json
src/pybind/mgr/prometheus/module.py