carlosedp / docker-arm_exporter Goto Github PK

View Code? Open in Web Editor NEW

20.0 5.0 12.0 20 KB

ARM CPU temperature exporter for Prometheus - Container builder

Shell 28.12% Dockerfile 71.88%

docker-arm_exporter's Introduction

docker-arm_exporter

ARM CPU temperature exporter for Prometheus. Exporter code by Lukas Malkmus

Usage:

Docker Run

Run the container in the same network as your Prometheus

docker run -d \
  --name=arm_exporter \
  --restart always \
  --net=monitoring \
  -p 9243:9243 \
  -v /etc/hostname:/etc/nodename:ro \
  -v /etc/localtime:/etc/localtime:ro \
  -v /etc/timezone:/etc/TZ:ro \
  carlosedp/arm_exporter

Docker Compose File (yml)

rpi-exporter:
    image: carlosedp/arm_exporter
    environment:
      - NODE_ID={{.Node.ID}}
    volumes:
      - /etc/hostname:/etc/nodename:ro
      - /etc/localtime:/etc/localtime:ro
      - /etc/timezone:/etc/TZ:ro
    command:
      - '--collector.textfile.directory=/etc/rpi_exporter/'
    ports:
      - 9243:9243
    networks:
      - monitoring
    deploy:
      mode: global

networks:
  monitoring:
    driver: overlay

Docker Stack

docker stack deploy monitoring --compose-file docker-compose-monitoring.yml

Add target to Prometheus:

- job_name: 'node'
  scrape_interval: 10s
  dns_sd_configs:
  - names:
    - 'arm_exporter'
    type: 'A'
    port: 9243

Get images on DockerHub

docker-arm_exporter's People

Contributors

Stargazers

Watchers

Forkers

jessestuart geekdave retocode akolk huntipl mhansen rossbachp papodaca solarnz dschunack bhuism encryptic

docker-arm_exporter's Issues

Latest version in DockerHub (published Sept 10, 2021) fails to start

This release is broken, constantly failing with:

/bin/rpi_exporter: line 1: syntax error: unexpected "|"

This is happening on both armv7 and arm64

Update to v0.6.0

Hi Carlos,
could you update to v0.6.0? Got some changes which might be worth updating: https://github.com/lukasmalkmus/rpi_exporter/releases

Build tagged releases of the exporter

This will allow folks to upgrade to a specific release as desired vs potentially running a mix of releases.

rpi_exporter: error: unexpected /bin/rpi_exporter, try --help

So I think CMD needs to be empty, since entrypoint is already calling /bin/rpi_exporter

/usr/bin/docker run --rm -p 9243:9243 -v /etc/hostname:/etc/nodename:ro -v /etc/localtime:/etc/localtime:ro -v /etc/timezone:/etc/TZ:ro --name=arm-exporter.service carlosedp/arm_exporter
rpi_exporter: error: unexpected /bin/rpi_exporter, try --help

 /usr/bin/docker run --rm -p 9243:9243 -v /etc/hostname:/etc/nodename:ro -v /etc/localtime:/etc/localtime:ro -v /etc/timezone:/etc/TZ:ro --entrypoint "/bin/sh" --name=arm-exporter.service carlosedp/arm_exporter -x /etc/rpi_exporter/docker-entrypoint.sh /bin/rpi_exporter

+ '[' -f /etc/nodename ]
+ cat /etc/nodename
+ NODE_NAME=raspberrypi
+ echo 'node_meta{node_id="none", container_label_com_docker_swarm_node_id="none", node_name="raspberrypi"} 1'
+ set -- /bin/rpi_exporter /bin/rpi_exporter
+ exec /bin/rpi_exporter /bin/rpi_exporter
rpi_exporter: error: unexpected /bin/rpi_exporter, try --help

Error on GPU collector

Hi,

 k logs -n monitoring arm-exporter-rpi-exporter-c5wjq
 level=error msg="gpu collector failed after 0.053595s: exit status 255" source="collector.go:142"

this has been fixed on the upstream project via lukasmalkmus/rpi_exporter#18.
Could you please update the docker image?
That would be awesome!

# for arm64 based installations
command:
        - /bin/rpi_exporter
        args:
          - --vcgencmd=/usr/bin/vcgencmd

Can't start on armv6 (Raspberry Pi Zero W)

Hey there, looks like 52d0771 removed support for arm32v6, replacing it with linux/arm,linux/arm64.

I've been using this container for a while on Raspberry Pi Zero W and it's started crashlooping with Restarting (139) 7 seconds ago, probably because of an unrecognized ARMv7 instruction.

I don't suppose this would be as simple to fix as adding linux/arm32v6 like this?

    - ARCHS="linux/arm,linux/arm64"
    - ARCHS="linux/arm,linux/arm64,linux/arm/v6"

Looks like linux/arm/v6 is the format according to https://www.docker.com/blog/getting-started-with-docker-for-arm-on-linux/

Btw, thanks for maintaining this, and no worries if you're not interested in supporting ARMv6, I can do a fork - I've gotten a lot of value already out of your image :-)

Can't start on armv6 (Raspberry Pi Zero W)

Hi there, looks like #9 is back. I'm now seeing this on my Raspberry Pi Zero W:

pi@zerow ~> docker ps
CONTAINER ID        IMAGE                                COMMAND                  CREATED             STATUS                            PORTS                    NAMES
34e224458e48        carlosedp/arm_exporter               "/etc/rpi_exporter/d…"   22 minutes ago      Restarting (139) 34 seconds ago                            dc_arm_exporter_1

Continual restarting.

I suspect it's only being built for ARMv7:

https://hub.docker.com/r/carlosedp/arm_exporter/tags?page=1&ordering=last_updated lists only linux/arm/v7 and linux/arm64.

I'm guessing this broke with the recent move from Travis to GitHub actions, here: f07e1ce#diff-5b21991be47c2922383bdc0b6bf00b65af7db51b82d049dd8d6ad03e3d37ac98L66, where

          platforms: linux/arm/v6,linux/arm/v7,linux/arm64

became:

          platforms: linux/arm,linux/arm64

Maybe we could revert that line? But I'm not sure, maybe was a reason that line had to be changed.

I made a pull request in the past: #10, maybe we could apply the same tricks in there to the new CI?

I installed the cluster monitoring on my k3s cluster that is in another one of your repos and it's awesome. The only issue was that prometheus was consuming a lot of memory on the cluster, by far more than another other pod. I really love the monitoring but not willing to give up all of that memory for metrics. I found that grafana now has a new agent, paired with a free account I can now collect metrics along with kube-state-metrics. It's collector runs at 25% of prometheus and seems to work really well. Yet I'm just missing the arm exporter so I can monitor and be alerted on temp issues.

https://grafana.com/products/cloud/

I have attempted the follow scrape configs with no avail, I simply cannot get it to scape my arm-exporter pods...

Any help would be appreciated.

scrape_configs:
        - job_name: 'rpi-nodes'
          scrape_interval: 10s
          dns_sd_configs:
          - names:
            - 'arm_exporter'
            type: 'A'
            port: 9243
        - job_name: kubernetes-pods
          kubernetes_sd_configs:
            - role: pod
          relabel_configs:
            - source_labels: [__meta_kubernetes_pod_annotation_prometheus_io_scrape]
              action: keep
              regex: true
            - source_labels: [__meta_kubernetes_pod_annotation_prometheus_io_path]
              action: replace
              target_label: __metrics_path__
              regex: (.+)
            - source_labels: [__address__, __meta_kubernetes_pod_annotation_prometheus_io_port]
              action: replace
              regex: ([^:]+)(?::\d+)?;(\d+)
              replacement: $1:$2
              target_label: __address__
            - action: labelmap
              regex: __meta_kubernetes_pod_label_(.+)
            - source_labels: [__meta_kubernetes_namespace]
              action: replace
              target_label: namespace
            - source_labels: [__meta_kubernetes_pod_name]
              action: replace
              target_label: pod

While also updating my daemonset:

apiVersion: apps/v1
kind: DaemonSet
metadata:
  labels:
    k8s-app: arm-exporter
  name: arm-exporter
  namespace: grafana
spec:
  selector:
    matchLabels:
      k8s-app: arm-exporter
  template:
    metadata:
      labels:
        k8s-app: arm-exporter
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "9243"
        prometheus.io/path: "/metrics"
...

/opt/vc/bin/vcgencmd: no such file or directory

The exporter running in a container produces the following errors:

time="2019-11-26T20:54:20-08:00" level=error msg="gpu collector failed after 0.007025s: fork/exec /opt/vc/bin/vcgencmd: no such file or directory" source="collector.go:142"

The process tries to execute the binary (lukasmalkmus/rpi_exporter#7) which is available on the host but not in the container. Is there a way to mount this binary and probably its dependencies into the container or it would require some changes in the exporter itself?