This reference blog documents every metric group collected by the IBM Storage Virtualize Management Pack. For each resource type, we list the metric key, its display name (from the English resource bundle), unit, data type, and how it is calculated. Use this as a definitive reference for understanding the data behind the dashboards and reports.

Unit conversions: Byte metrics (hbt, hbr, cbt, cbr, lnbt, lnbr, rmbt, rmbr) are DELTA calculated then divided by 1,048,576 to convert bytes → MB. Other metrics are either DELTA or LATEST as noted.

💿 Internal Drive Metrics

Metric Accuracy and Timestamp Notes

• Timestamp anchoring: The SystemCollector calls lssystemstats at the start of each cycle and caches the sample_epoch (epoch seconds → milliseconds). All collectors that lack their own file timestamps use this cached value as fallback — ensuring all metrics in a single cycle share a consistent time anchor.
• Calculated interval: For all stat file pipelines (Nv, Nh, Nn, Nm, Nd), the interval is calculated as the difference between the timestamps of the two downloaded files (new − old) in seconds, rather than using a hardcoded 300s constant. This handles cases where collection was delayed or the FlashSystem regenerated files at a non-standard interval.
• Counter rollover: For HostPort FPIN metrics (lin, pcn), if the new value is less than the old value (counter rollover), the adapter uses the new value directly rather than a negative delta.
• Multi-node aggregation: IBM FlashSystem generates one stats file per node (panel). The adapter downloads files from all nodes and merges them: cumulative counters are summed, while peak latency metrics (rlw, wlw, tmaxdelay) take the maximum value across nodes.
  

  
  For any queries feel free to contact:

  • Anushka Gupta @Anushka Gupta
  • Anmol Chaudhary @Anmol Chaudhary
  • Mayur Bhokase @Mayur Bhokase
  • Sandeep Bangur @SANDEEP BANGUR