Ceph¶

The ceph probe monitors a Ceph cluster via the Ceph REST Management API v1, reporting cluster health, OSD counts, monitor quorum, capacity and per-pool I/O statistics.

Quick start¶

# probes.d/10-ceph.yaml — each file under probes.d/ is a YAML array of probes
- name: ceph
  type: ceph
  params:
    endpoint: https://localhost:8443
    username: admin
    password: ${secret:ceph.password}   # OS secret store; inline plaintext is auto-sealed on install

Parameters¶

Parameter	Must set	Default	Description
`endpoint`	In practice	`https://localhost:8443`	Base URL of the Manager dashboard / REST API
`username`	Yes	-	Dashboard user
`password`	Yes	-	Dashboard user's password. A secret: reference it with `${secret:…}`, `${env:…}` or `${file:…}` rather than writing it in the file
`verify_tls`	No	`true`	Verify the dashboard certificate; false accepts a self-signed one
`interval`	No	`60`	Seconds between collections
`instance_name`	No	-	Stable identity override for this cluster

Metrics¶

Metric	Unit	Description
`senhub.ceph.up`	1	1 when the Ceph REST API is reachable and authentication succeeds
`ceph.health.status`	1	Cluster health: 2 = HEALTH_OK, 1 = HEALTH_WARN, 0 = HEALTH_ERR
`ceph.cluster.capacity`	By	Raw cluster storage capacity
`ceph.cluster.used`	By	Raw storage space in use
`ceph.osd.count`	{osd}	Total number of OSDs in the cluster
`ceph.osd.up`	{osd}	OSDs currently in the `up` state
`ceph.osd.in`	{osd}	OSDs currently in the `in` state (participating in data placement)
`ceph.monitor.quorum.count`	{monitor}	Number of monitors participating in the quorum
`ceph.pool.rd_ops`	{read}	Read operations per pool (tagged with `pool`)
`ceph.pool.wr_ops`	{write}	Write operations per pool

Operational notes¶

The Ceph Manager Dashboard must be enabled: ceph mgr module enable dashboard.
The dashboard ships with a self-signed TLS certificate, which the probe rejects by default. Set verify_tls: false to accept it in a lab; for production, configure a proper certificate.
The API requires Ceph Nautilus (14+) or newer for the /api/ v1 interface.

Metric reference¶

Every metric this probe can emit. Metric is the OpenTelemetry name the OTLP, Prometheus and Zabbix outputs derive theirs from. Name is what a Nagios check and the API metrics= filter match. PRTG channel is the label PRTG shows, placeholders filled from the series' tags.

Metric	Name	PRTG channel	Unit	Description
`senhub.ceph.up`	`senhub.ceph.up`	Ceph {instance} Reachability	#	1 when the Ceph REST API is reachable and authentication succeeds
`ceph.health.status`	`ceph.health.status`	Ceph {instance} Health Status	#	Cluster health: 2=HEALTH_OK, 1=HEALTH_WARN, 0=HEALTH_ERR
`ceph.cluster.capacity`	`ceph.cluster.capacity`	Ceph {instance} Raw Capacity	B	Total raw cluster capacity in bytes
`ceph.cluster.used`	`ceph.cluster.used`	Ceph {instance} Used Capacity	B	Total bytes currently used across the cluster
`ceph.osd.count`	`ceph.osd.total`	Ceph {instance} OSD Total	#	Total number of OSDs configured in the cluster
`ceph.osd.in`	`ceph.osd.in`	Ceph {instance} OSD In	#	Number of OSDs that are in (participating in the cluster)
`ceph.osd.up`	`ceph.osd.up`	Ceph {instance} OSD Up	#	Number of OSDs that are up (running)
`ceph.monitor.count`	`ceph.monitor.count`	Ceph {instance} Monitor Count	#	Total number of monitor daemons
`ceph.monitor.quorum.count`	`ceph.monitor.quorum_count`	Ceph {instance} Monitor Quorum	#	Number of monitors currently in quorum
`ceph.pool.objects`	`ceph.pool.objects`	Ceph {instance} Pool {pool} Objects	#	Number of objects stored in the pool
`ceph.pool.used`	`ceph.pool.used`	Ceph {instance} Pool {pool} Used	B	Bytes stored in the pool
`ceph.pool.rd_ops`	`ceph.pool.rd_ops`	Ceph {instance} Pool {pool} Read Ops	#	Cumulative read operations on the pool
`ceph.pool.wr_ops`	`ceph.pool.wr_ops`	Ceph {instance} Pool {pool} Write Ops	#	Cumulative write operations on the pool