Collecting metrics by using Prometheus exporters

You can use Prometheus exporters to collect metrics from hosts, services, or apps which do not natively expose Prometheus-formatted metrics. You can monitor these metrics through a Monitoring instance.

See the IBM support statement for the use of Prometheus exporters.

There are different sources for Prometheus exporters:

Official exporters are available in the official Prometheus GitHub organization and are labelled official.
Sysdig curates and maintains PromCat. PromCat is an enterprise resource catalog where you can find supported Monitoring integrations for Kubernetes platforms and cloud-native services.

You can collect metrics from different sources such as:

Hosts for which a Monitoring agent is not available such as Windows systems or VMware ESXi-Host systems.
Hosts for which a Monitoring agent is available, but you need to collect other types of metrics such as IPMI sensor metrics or hardware and kernel-relaed metrics.
Services like MySQL database

The following table lists some Prometheus exporters that you can use to monitor your infrastructure:

Exporters
Exporters	Use case	Source
`Blackbox Exporter`	Allows blackbox probing of endpoints over HTTP, HTTPS, DNS, TCP and ICMP. The monitoring agent can be used in conjunction with the Blackbox exporter to collect availability metrics.	Prometheus Blackbox exporter (official)
`IPMI Exporter`	Collects Intelligent Platform Management Interface (IPMI) device sensor metrics.	Prometheus IPMI exporter (opensource)
`Windows WMI Exporter`	Collects Windows system metrics.	PromCat: Windows Exporter (opensource)
`Node Exporter`	Collects hardware and kernel-related metrics that are exposed by *NIX kernels.	Node exporter (official)
`VMware Exporter`	Collects metrics from VMware vCenter deployments.	VMWare exporter (Opensource)

Windows exporter

Configure the Prometheus windows_exporter to collect Windows system metrics.

The Prometheus Windows exporter runs as a Windows service. You configure the metrics that you want to monitor by enabling collectors.

The following collectors are supported:

Collectors
Collector name	Information about metrics collected per collector
`cpu`	CPU metrics
`cs`	Computer system metrics
`logical_disk`	Disk metrics
`os`	Operating System metrics
`system`	System metrics
`net`	Network interface metrics
`memory`	Memory metrics

To learn how to configure the Windows exporter, see Monitoring a Windows environment.

The legacy Prometheus WMI Exporter is still supported, see Monitoring a Windows environment using the legacy WMI Exporter for information on the WMI Exporter.

IPMI exporter

In addition to the set of metrics that are automatically collected by the monitoring agent, you might want to collect other metrics such as sensor metrics.

Configure the Prometheus IPMI exporter to collect Intelligent Platform Management Interface (IPMI) device sensor metrics.

The Prometheus IPMI Exporter exporter supports local IPMI devices and remote devices that can be accessed by using Remote Management Control Protocol (RMCP).
When you use RMCP to access remote devices, you can use an IPMI exporter to monitor multiple IPMI devices. You identify each device by passing the target hostname as a parameter.
The IPMI exporter relies on tools from the FreeIPMI suite.

You can collect the following metrics when you configure the IPMI exporter:

IPMI admin metrics

The metric ipmi_up {collector="<NAME>"} reports 1 when data from a remote host is collected successfully. It reports 0 for collection of data in a local host.

The metric ipmi_scrape_duration_seconds reports the amount of time that it takes the collector to retrieve the data.
IPMI System event log (SEL) metrics

The metric ipmi_sel_entries_count reports the number of entries in the system event log.

The metric ipmi_sel_free_space_bytes reports the number of free bytes for new ystem event log entries.
IPMI sensor data

The IPMI exporter collects 2 metrics per sensor type: state and value. A value of 0 reports a normal state. A value of 1 reports a warning state. A value of 2 reports a critical state. A value of NaN reports information not available. For example, see the metrics for different sensors:

Temperature sensor metrics: ipmi_temperature_celsius, ipmi_temperature_state

Fan speed sensor metrics: ipmi_fan_speed_rpm, ipmi_fan_speed_state

Voltage sensor metrics: ipmi_voltage_state, ipmi_voltage_volts
IPMI chassis power state of the machine

The metric ipmi_chassis_power_state informs about the current state of the chassis of the machine. It has a value of 1 when the power is on. It has a value of 0 when the power is off.
DCMI data

The metric ipmi_dcmi_power_consumption_current_watts informs about the live power consumption of the machine in Watts.
BMC details

The metric ipmi_bmc_info includes information about the firmware revision and manufacturer in labels and has a value of 1.

For more information, see Prometheus IPMI Exporter.

To learn how to configure the IPMI exporter, see Configuring the Prometheus IPMI Exporter to monitor sensor metrics.

You can also check the tutorial: Configure the Prometheus IPMI Exporter to monitor sensor metrics in a Bare metal.

Blackbox exporter

Allows blackbox probing of endpoints over HTTP, HTTPS, DNS, TCP and ICMP. The monitoring agent can be used in conjunction with the Blackbox exporter to collect availability metrics.

Configure the Prometheus Blackbox exporter to monitor host availability, such as URL sites.

Node exporter

In addition to the set of metrics that are automatically collected by the monitoring agent, you might want to collect other Linux host metrics that are exposed by *NIX kernels.

Configure the Prometheus Node exporter to collect hardware and kernel metrics that are exposed by *NIX kernels.

For more information, see Monitoring Linux host metrics with the Node Exporter.

VMWare exporter

For information on configuring a VMWare exporter, see Monitoring for VMware vCenter Server deployments