Skip to main content

Commonly used Prometheus metrics

Contributors netapp-lhalbert

Refer to this list of commonly used Prometheus metrics to better understand conditions in the default alert rules or to construct the conditions for custom alert rules.

For details on the syntax of Prometheus queries, see Querying Prometheus.

What are Prometheus metrics?

Prometheus metrics are time series measurements. The Prometheus service on Admin Nodes collects these metrics from the services on all nodes. Metrics are stored on each Admin Node until the space reserved for Prometheus data is full. When the /var/local/mysql_ibdata/ volume reaches capacity, the oldest metrics are deleted first.

Where are Prometheus metrics used?

The metrics collected by Prometheus are used in several places in the Grid Manager:

  • Nodes page: The graphs and charts on the tabs available from the Nodes page use the Grafana visualization tool to display the time-series metrics collected by Prometheus. Grafana displays time-series data in graph and chart formats, while Prometheus serves as the backend data source.

    Prometheus graph
  • Alerts: Alerts are triggered at specific severity levels when alert rule conditions that use Prometheus metrics evaluate as true.

  • Grid Management API: You can use Prometheus metrics in custom alert rules or with external automation tools to monitor your StorageGRID system. A complete list of Prometheus metrics is available from the Grid Management API. (From the top of the Grid Manager, select the help icon and select API documentation > metrics.) While more than a thousand metrics are available, only a relatively small number are required to monitor the most critical StorageGRID operations.

    Note Metrics that include private in their names are intended for internal use only and are subject to change between StorageGRID releases without notice.
  • The SUPPORT > Tools > Diagnostics page and the SUPPORT > Tools > Metrics page: These pages, which are primarily intended for use by technical support, provide several tools and charts that use the values of Prometheus metrics.

    Note Some features and menu items within the Metrics page are intentionally non-functional and are subject to change.

List of most common metrics

The following list contains the most commonly used Prometheus metrics.

Note Metrics that include private in their names are for internal use only and are subject to change without notice between StorageGRID releases.
alertmanager_notifications_failed_total

The total number of failed alert notifications.

node_filesystem_avail_bytes

The amount of file system space available to non-root users in bytes.

node_memory_MemAvailable_bytes

Memory information field MemAvailable_bytes.

node_network_carrier

Carrier value of /sys/class/net/iface.

node_network_receive_errs_total

Network device statistic receive_errs.

node_network_transmit_errs_total

Network device statistic transmit_errs.

storagegrid_administratively_down

The node is not connected to the grid for an expected reason. For example, the node, or services on the node, has been gracefully shut down, the node is rebooting, or the software is being upgraded.

storagegrid_appliance_compute_controller_hardware_status

The status of the compute controller hardware in an appliance.

storagegrid_appliance_failed_disks

For the storage controller in an appliance, the number of drives that aren't optimal.

storagegrid_appliance_storage_controller_hardware_status

The overall status of the storage controller hardware in an appliance.

storagegrid_content_buckets_and_containers

The total number of S3 buckets and Swift containers known by this Storage Node.

storagegrid_content_objects

The total number of S3 and Swift data objects known by this Storage Node. Count is valid only for data objects created by client applications that interface with the system through S3.

storagegrid_content_objects_lost

The total number of objects this service detects as missing from the StorageGRID system. Action should be taken to determine the cause of the loss and if recovery is possible.

storagegrid_http_sessions_incoming_attempted

The total number of HTTP sessions that have been attempted to a Storage Node.

storagegrid_http_sessions_incoming_currently_established

The number of HTTP sessions that are currently active (open) on the Storage Node.

storagegrid_http_sessions_incoming_failed

The total number of HTTP sessions that failed to complete successfully, either due to a malformed HTTP request or a failure while processing an operation.

storagegrid_http_sessions_incoming_successful

The total number of HTTP sessions that have completed successfully.

storagegrid_ilm_awaiting_background_objects

The total number of objects on this node awaiting ILM evaluation from the scan.

storagegrid_ilm_awaiting_client_evaluation_objects_per_second

The current rate at which objects are evaluated against the ILM policy on this node.

storagegrid_ilm_awaiting_client_objects

The total number of objects on this node awaiting ILM evaluation from client operations (for example, ingest).

storagegrid_ilm_awaiting_total_objects

The total number of objects awaiting ILM evaluation.

storagegrid_ilm_scan_objects_per_second

The rate at which objects owned by this node are scanned and queued for ILM.

storagegrid_ilm_scan_period_estimated_minutes

The estimated time to complete a full ILM scan on this node.

Note: A full scan does not guarantee that ILM has been applied to all objects owned by this node.

storagegrid_load_balancer_endpoint_cert_expiry_time

The expiration time of the load balancer endpoint certificate in seconds since the epoch.

storagegrid_metadata_queries_average_latency_milliseconds

The average time required to run a query against the metadata store through this service.

storagegrid_network_received_bytes

The total amount of data received since installation.

storagegrid_network_transmitted_bytes

The total amount of data sent since installation.

storagegrid_node_cpu_utilization_percentage

The percentage of available CPU time currently being used by this service. Indicates how busy the service is. The amount of available CPU time depends on the number of CPUs for the server.

storagegrid_ntp_chosen_time_source_offset_milliseconds

Systematic offset of time provided by a chosen time source. Offset is introduced when the delay to reach a time source is not equal to the time required for the time source to reach the NTP client.

storagegrid_ntp_locked

The node is not locked to a Network Time Protocol (NTP) server.

storagegrid_s3_data_transfers_bytes_ingested

The total amount of data ingested from S3 clients to this Storage Node since the attribute was last reset.

storagegrid_s3_data_transfers_bytes_retrieved

The total amount of data retrieved by S3 clients from this Storage Node since the attribute was last reset.

storagegrid_s3_operations_failed

The total number of failed S3 operations (HTTP status codes 4xx and 5xx), excluding those caused by S3 authorization failure.

storagegrid_s3_operations_successful

The total number of successful S3 operations (HTTP status code 2xx).

storagegrid_s3_operations_unauthorized

The total number of failed S3 operations that are the result of an authorization failure.

storagegrid_servercertificate_management_interface_cert_expiry_days

The number of days before the Management Interface certificate expires.

storagegrid_servercertificate_storage_api_endpoints_cert_expiry_days

The number of days before the Object Storage API certificate expires.

storagegrid_service_cpu_seconds

The cumulative amount of time that the CPU has been used by this service since installation.

storagegrid_service_memory_usage_bytes

The amount of memory (RAM) currently in use by this service. This value is identical to that displayed by the Linux top utility as RES.

storagegrid_service_network_received_bytes

The total amount of data received by this service since installation.

storagegrid_service_network_transmitted_bytes

The total amount of data sent by this service.

storagegrid_service_restarts

The total number of times the service has been restarted.

storagegrid_service_runtime_seconds

The total amount of time that the service has been running since installation.

storagegrid_service_uptime_seconds

The total amount of time the service has been running since it was last restarted.

storagegrid_storage_state_current

The current state of the storage services. Attribute values are:

  • 10 = Offline

  • 15 = Maintenance

  • 20 = Read-only

  • 30 = Online

storagegrid_storage_status

The current status of the storage services. Attribute values are:

  • 0 = No Errors

  • 10 = In Transition

  • 20 = Insufficient Free Space

  • 30 = Volume(s) Unavailable

  • 40 = Error

storagegrid_storage_utilization_data_bytes

An estimate of the total size of replicated and erasure-coded object data on the Storage Node.

storagegrid_storage_utilization_metadata_allowed_bytes

The total space on volume 0 of each Storage Node that is allowed for object metadata. This value is always less than the actual space reserved for metadata on a node, because a portion of the reserved space is required for essential database operations (such as compaction and repair) and future hardware and software upgrades.The allowed space for object metadata controls overall object capacity.

storagegrid_storage_utilization_metadata_bytes

The amount of object metadata on storage volume 0, in bytes.

storagegrid_storage_utilization_total_space_bytes

The total amount of storage space allocated to all object stores.

storagegrid_storage_utilization_usable_space_bytes

The total amount of object storage space remaining. Calculated by adding together the amount of available space for all object stores on the Storage Node.

storagegrid_swift_data_transfers_bytes_ingested

The total amount of data ingested from Swift clients to this Storage Node since the attribute was last reset.

storagegrid_swift_data_transfers_bytes_retrieved

The total amount of data retrieved by Swift clients from this Storage Node since the attribute was last reset.

storagegrid_swift_operations_failed

The total number of failed Swift operations (HTTP status codes 4xx and 5xx), excluding those caused by Swift authorization failure.

storagegrid_swift_operations_successful

The total number of successful Swift operations (HTTP status code 2xx).

storagegrid_swift_operations_unauthorized

The total number of failed Swift operations that are the result of an authorization failure (HTTP status codes 401, 403, 405).

storagegrid_tenant_usage_data_bytes

The logical size of all objects for the tenant.

storagegrid_tenant_usage_object_count

The number of objects for the tenant.

storagegrid_tenant_usage_quota_bytes

The maximum amount of logical space available for the tenant's objects. If a quota metric is not provided, an unlimited amount of space is available.

Get a list of all metrics

To obtain the complete list of metrics, use the Grid Management API.

  1. From the top of the Grid Manager, select the help icon and select API documentation.

  2. Locate the metrics operations.

  3. Execute the GET /grid/metric-names operation.

  4. Download the results.