Skip to content

retzkek/chiamon

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ChiaMon

Example Chia monitoring stack, using:

This includes a docker-compose configuration to run everything, but this is primarily intended for development and testing.

WARNING this is NOT a one-click install, expect to need to do some work setting everything up for your machine. PLEASE read the notes below and understand what all the services are, what they do, and how they work together.

Chia dashboard

mtail program

The mtail program is in mtail/chialog.mtail. Currently it only collects harvester metrics:

  • chia_harvester_blocks_total: cumulative number of block challenges attempted
  • chia_harvester_plots_total: current number of plots
  • chia_harvester_plots_eligible: cumulative number of plots that passed filter
  • chia_harvester_proofs_total: cumulative number of proofs won
  • chia_harvester_search_time: histogram of proof search times

NOTE you need to set log_level to INFO in your Chia config.yaml to get harvester metrics.

chia_exporter

The chia_exporter is used to collect metrics from the Chia node RPC API.

Grafana dashboard

The example Grafana dashboard is in grafana/dashboards/Chia.json. It defines a number of variables that will be auto-populated from the node metrics. Grafana dashboards are easily customized to show what you're interested in seeing, in the way you find best; this dashboard is just meant to demonstrate what can be done.

Running on Linux/Mac

The docker-compose file will mount the Chia log from $HOME/.chia/mainnet/log/debug.log, verify that this location is correct and set the log level to INFO in the Chia configuration (usually at $HOME/.chia/mainnet/config/config.yaml).

Run:

docker-compose up -d

This will do the following:

  • Build container image with configuration for mtail from source
  • Build container image for chia_exporter from source
  • Download other images from docker hub
  • Run containers in the background, attached to the host network (this makes it easy to communicate with native services, but has some trade-offs. See notes.)

The grafana service provisions the prometheus and loki datasources and a basic dashboard that displays harvester and node metrics.

Access Grafana at http://localhost:3000 and login with the default admin/admin username and password (you'll be prompted to change the password).

Notes

  • It's highly encouraged to run the node exporter natively rather than in docker - see the discussion in the node_exporter docs. On Ubuntu you can run sudo apt install prometheus-node-exporter, which includes disk SMART monitoring (disk temperatures, etc) as well. If you do run it in Docker, you'll need to bind-mount in any other volumes you want to monitor (add them to the volumes list in docker-compose.yml, e.g. - '/scratch:/scratch'). See issue #3.

  • The docker-compose file uses the $HOME environment variable for the Chia log paths. Verify that these paths are correct, and if you run the docker-compose commands with sudo then you'll have to replace $HOME with the actual path (since root's home is not your home!). Even better, add your user to the docker group so you don't have to use sudo:

      sudo usermod -a -G docker username
    
  • On Mac you'll need to run node_exporter natively, not under Docker: brew install node_exporter. You'll probably need to change the networking setup too, since Docker on Mac runs in a VM. See the windows docker-compose and prometheus configs.

Running on Windows

The node exporter does not work on Windows; instead you need to use the Windows exporter for system metrics. Modified config and example dashboard are in the windows branch. You may also want to review the discussion in issue #2.

These steps will get you to a working setup (but aren't the only way):

Monitoring Multiple Nodes

To monitor multiple nodes (e.g. multiple harvesters), you just need to run the appropriate exporters (e.g. node_exporter and mtail for a harvester) and add them as targets in the prometheus config, for example:

  - job_name: 'node'
    static_configs:
      - targets: ['localhost:9100', 'harvester1:9100', 'harvester2:9100']
  - job_name: 'mtail'
    static_configs:
      - targets: ['localhost:3903', 'harvester1:3903', 'harvester2:3903']

If you're also running Loki to collect logs, you'll also want to run promtail on every node, and configure it to push logs to Loki on your monitoring node:

clients:
  - url: http://loki-server:3100/loki/api/v1/push

Copyright & License

Copyright 2021 Kevin Retzke

This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details.

See LICENSE.txt