Number of volumes

[nixpanic]

Number of volumes

[shtripat]

No of volumes is more cluster level metric. In additional below would be useful metrics

• no of bricks
• no of gluster pods (no of peers)
• cluster overall status (healthy/unhealthy)
• status wise split of volumes (e.g. UP 2, DOWN 3)
• status wise split of bricks (e.g. Started 2, Stopped 1)
• device count
• overall capacity and utilization

i think this can be achieved from prometheus recording rules using count_over_time()

[shtripat]

No of volumes is more cluster level metric. In additional below would be useful metrics

* no of bricks

Done

* no of gluster pods (no of peers)

Will be taken care by k8s

* cluster overall status (healthy/unhealthy)

Operator to tackle this as operator itself needs to be monitored and should have some metrics exposed. Cluster overall health is ideal candidate for this.

* status wise split of volumes (e.g. UP 2, DOWN 3)

A metric gluster_volume_up exposed and recording rules could be used for total no of up/down volumes required.

* status wise split of bricks (e.g. Started 2, Stopped 1)

A metric gluster_brick_up exposed and recording rule could be used for total no of up/down bricks if required.

* device count

k8s exposes the device specific metrics and recording rules could be used for getting the overall count

* overall capacity and utilization

Done.

Based on above comment, I am closing this issue. Kindly raise separate ones on need basis.