Cinder Volume Performance and Health Stats Collection

Enable the collection of block storage driver volume performance and health value collection for use in ceilometer and horizon.

Extend the current cinder-volume driver interface and cinder volume manager to support per-volume performance collection, per-volume health status, and driver instance health status (per array health).

The initial volume performance metrics will be as follows:

- “Read IOPS (IO/s)” – ‘iops_read’
- “Write IOPS (IO/s)” – ‘iops_write’
-“Read Throughput (KB/s)” – ‘throughput_read’
- “Write Throughput (KB/s)” – ‘throughput_write’
- “Read Latency (us)” – ‘latency_read’
- “Write Latency (us)” – ‘latency_write’

The initial volume health metrics will be as follows:
- “Vendor Volume Health” – ‘vendor_health’ – ‘OPERATIONAL’|’WARNING’|’CRITICAL’
- “Vendor Volume Substatus” - ‘vendor_substatus’ – a vendor specific field to show status such as ‘Pool health degraded’, ‘performance throttled due to QOS settings’, …

The initial instance health metrics will be as follows:
- “Array Health Status” – ‘array_status’
- “Array Health SubStatus” – ‘array_substatus’

If the cinder volume driver supports the ‘perf_collection’ feature the cinder volume manager would define and add new periodic stats
- Using the ‘@periodic_task decorator, every 15 seconds call a new cinder volume driver interface get_volume_performance_stats()
- Using the ‘@periodic_task decorator, every 60 seconds call a new cinder volume driver interface get_volume_health_stats()
- The existing get_volume_stats() call would have two new attributes added to its expected response object:
o ‘array_status’
o ‘array_substatus’