HP iLO Metrics Exporter

Blackbox like exporter used to export HP Server Integrated Lights Out (iLO) metrics to Prometheus. ILO version 1 to 4 is supported.

Gauges

Here are the status code of gauge

0 - OK
1 - Degraded
2 - Dead (Other)

Output example

hpilo-exporter is using [hpilo] (https://seveas.github.io/python-hpilo/) to retrieve health information. A typical reponse from a health_at_a_glance call will look like:

health_at_a_glance:
  battery: {status: OK}
  bios_hardware: {status: OK}
  fans: {redundancy: Redundant, status: OK}
  memory: {status: OK}
  network: {status: Link Down},
  power_supplies: {redundancy: Redundant, status: OK}
  processor: {status: OK}
  storage: {status: Degraded}
  temperature: {status: OK}
  vrm: {status: Ok}
  drive: {status: Ok}

The returned output would be:

hpilo_battery{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_storage{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 1.0
hpilo_fans{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_bios_hardware{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_memory{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_power_supplies{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_processor{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_network{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 2.0
hpilo_temperature{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 0.0
hpilo_vrm{product_name="ProLiant DL380 Gen6",server_name="name.fqdn.domain"} 0.0
hpilo_drive{product_name="ProLiant DL380 Gen6",server_name="name.fqdn.domain"} 0.0
hpilo_firmware_version{product_name="ProLiant DL360 Gen9",server_name="name.fqdn.domain"} 2.5

Installing

You can install exporter on the server directly or on separate machine. To run, you must have Python and pip installed.

To install with pip:

pip install -e $HPILO_EXPORTER_DIR

Then just:

hpilo-exporter [--address=0.0.0.0 --port=9416 --endpoint="/metrics"]

Requesting metrics

To request metrics you have to provide the ilo target. This can be done by providing query arguments or by setting environemnt varaibles.

A request with query arguments will look like:

curl 'http://127.0.0.1:9416/metrics?ilo_host=127.0.0.1&ilo_port=443&ilo_user=admin&ilo_password=admin'

To specify the target using environment variables you have to define:

ILO_HOST=127.0.0.1
ILO_PORT=443
ILO_USER=admin
ILO_PASSWORD=admin

Now you can omit the query arguments on the metrics endpoint:

curl 'http://127.0.0.1:9416/metrics'

Any query argument will override the default provided on the environment.

Caching

Requesting metrics from iLO is not exactly the fastet operation and prometheus may timeout when trying to scrape the metrics endpoint. To work around this issue hpilo-exporter offers a cache mode.

With the cache mode activated the first request will trigger requesting metrics from ilo in the background and you will receive an error response. As soon as the first background metrics collection completes, hpilo-exporter will answer with the last operations result and schedules a new background metrics collection if not already in progress.

To enable the cache mode you can pass

ilo_cached=true

with your query parameters or set the environment variable

ILO_CACHED=true

Docker

Prebuild images are available from the docker repository:

idnt/hpilo-exporter:latest

To build the image yourself

docker build --rm -t hpilo-exporter .

To run the container

docker run -p 9416:9416 hpilo-exporter:latest

Passing argument to the docker run command:

docker run -p 9416:9416 hpilo-exporter:latest --port 9416 --ilo_user my_user --ilo_password my_secret_password

Docker compose

Here is an example of Docker Compose deployment:

hpilo:
    image: my.registry/hpilo-exporter
    ports:
      - 9416:9416
    command:
      - '--port=9416'
    deploy:
      placement:
        constraints:
          - node.hostname == my_node.domain

Kubernetes

A helm chart designed for openstack-helm deployments is available at prometheus-helm-addons.

Prometheus config

Assuming:

the exporter is available on http://hpilo:9416
you use same the port,username and password for all your iLO

- job_name: 'hpilo'
  scrape_interval: 1m
  params:
    ilo_port: ['443']
    ilo_user: ['my_ilo_user']
    ilo_password: ['my_ilo_password']
  static_configs:
    - targets:
      - ilo_fqdn.domain

  relabel_configs:
    - source_labels: [__address__]
      target_label: __param_ilo_host
    - source_labels: [__param_ilo_host]
      target_label: ilo_host
    - target_label: __address__
      replacement: hpilo:8082  # hpilo exporter.

IDNT / hpilo-exporter