About

Storj exporter for Prometheus written in python. It pulls information from storj node api for node, satellite and payout metrics.

Also check out Storj-Exporter-dashboard for Grafana to visualise metrics for multiple storj nodes.

Tested with storj node versions listed under tests/api_mock/

Usage

• Exporter can be installed as a docker container or a systemd service or a standalone script
• Make sure you have -p 127.0.0.1:14002:14002 in storagenode container docker run command to allow local connections to storj node api
• storagenode is the default value for STORJ_HOST_ADDRESS environment variable that sets the address of the storage node container used to link exporter to the api
• If you storagenode container has a different name it needs to be set with both --link=<storagenode-name-here> and -e STORJ_HOST_ADDRESS=<storagenode-name-here> on the docker command

Installation

Docker installation

Run latest build from DockerHub (easiest option, assuming storagenode is the name of the storagenode container)

docker run -d --link=storagenode --name=storj-exporter -p 9651:9651 -e STORJ_HOST_ADDRESS=storagenode anclrii/storj-exporter:latest

Docker image supports linux/386,linux/amd64,linux/arm/v6,linux/arm/v7,linux/arm64 platforms.

Run multiple instances of exporter to monitor multiple storagenodes running on the same host

In this example storagenode1, storagenode2, storagenode3 are the names of storagenode containers runnin on the same host. The docker commands would be:

docker run -d --link=storagenode1 --name=storj-exporter1 -p 9651:9651 -e STORJ_HOST_ADDRESS=storagenode1 anclrii/storj-exporter:latest
docker run -d --link=storagenode2 --name=storj-exporter2 -p 9652:9651 -e STORJ_HOST_ADDRESS=storagenode2 anclrii/storj-exporter:latest
docker run -d --link=storagenode3 --name=storj-exporter3 -p 9653:9651 -e STORJ_HOST_ADDRESS=storagenode3 anclrii/storj-exporter:latest

Systemd service installation

Create storj-exporter user for service

useradd --no-create-home --shell /bin/false storj_exporter

Install package dependencies

Dependencies: python3 python3-pip
pip3 install --no-cache-dir -r /requirements.txt

Move storj_exporter directory to a desired location

mv storj_exporter/ /opt/
chown storj_exporter:storj_exporter /opt/storj_exporter/

Install systemd service and set to start on boot

cp storj_exporter.service /etc/systemd/system/
systemctl daemon-reload
systemctl restart storj_exporter
systemctl enable storj_exporter

Standalone script

python3 /path/to/storj_exporter/

Installing full monitoring stack (Prometheus + Grafana + Dashboard)

You can find some installation notes and guides in dashboard README, also see quick-start guide to set up the whole stack using docker-compose.

Variables

Following environment variables are available:

Collectors

By default exporter collects node, payout and satellite data from api. Satellite data is particularly expensive on cpu resources and disabling it might be useful on smaller systems

Netdata

For users that use Netdata: Netdata by default has a prometheus plugin enabled, which pulls all the data from the exporter every 5 seconds. This results in high CPU spikes on the storagenode. It is therefore advisable to disable the prometheus plugin of Netdata:

cd /etc/netdata
sudo ./edit-config go.d.conf

Then under "modules:" uncomment "prometheus" and change its value to "no":

modules:
#  activemq: yes
[...]
#  powerdns_recursor: yes
  prometheus: no

After that restart the netdata service:

sudo systemctl restart netdata