Skip to main content

Puppetboard

PyPI Version Tests Status

This is a fork of voxpupuli/puppetboard by Vox Pupuli with a read-only JSON API, an MCP server for AI assistants, current dependencies and a hardened container image.

Puppetboard is a web interface to PuppetDB aiming to replace the reporting functionality of Puppet Enterprise console (previously: Puppet Dashboard) for the open source Puppet.

Overview

See more screenshots here.

Table of Contents

Requirements

  • PuppetDB v. 5.2-8.*
    • (Note: PuppetDB 8.1.0 is not supported because of this bug. Please update to 8.1.1+.)
  • Python 3.12-3.15 or Docker

Installation

This fork is available on PyPI as puppetboard-dp and as container image ghcr.io/dennispohlmann/puppetboard.

Using Docker

The container image of this fork is published in the GitHub Container Registry as ghcr.io/dennispohlmann/puppetboard.

You can run the app on your PuppetDB host with this command:

docker run -it \
  -e PUPPETDB_HOST=localhost \
  -e PUPPETDB_PORT=8080 \
  -e PUPPETBOARD_PORT=8088 \
  -e SECRET_KEY=XXXXXXXX \
  --net=host \
  ghcr.io/dennispohlmann/puppetboard

With --net=host next to PuppetDB, PUPPETBOARD_PORT must differ from PuppetDB's port 8080. Note: you must provide a secret key! Generate one for example by running ruby -e "require 'securerandom'; puts SecureRandom.hex(32)".

The image is based on the official python:3.15-alpine image, pinned by digest, without compiler and pip. It runs as the unprivileged user puppetboard (uid 10001) and listens on port 8080 (PUPPETBOARD_PORT).

Optionally you can set PUPPETBOARD_URL_PREFIX env variable to a value like /puppetboard to run the app under a URL prefix.

Ready-made docker-compose examples cover Puppetboard alone and behind nginx with TLS and Basic-Auth. For Kubernetes there is a Helm chart, released as oci://ghcr.io/dennispohlmann/charts/puppetboard.

The image can also serve a read-only JSON API (-e ENABLE_API=true) and an MCP server for AI assistants (-e ENABLE_MCP=true -e MCP_ALLOWED_HOSTS=puppetboard.example.com). Both have no authentication of their own, so put a TLS reverse proxy with authentication in front. See docs/API.md.

You can use the following Puppet Code to have Puppetboard managed by Puppet:

include docker

docker::image { 'ghcr.io/dennispohlmann/puppetboard': }

docker::run { 'puppetboard':
  image => 'ghcr.io/dennispohlmann/puppetboard',
  env   => [
    'PUPPETDB_HOST=127.0.0.1',
    'PUPPETDB_PORT=8080',
    'PUPPETBOARD_PORT=8088',
    'SECRET_KEY=XXXXXXXX',
  ],
  net   => 'host',
}

If you want to have all features enabled, you must use SSL talking to PuppetDB:

file { '/etc/puppetboard':
  ensure => directory,
}
file { '/etc/puppetboard/key.pem':
  ensure => file,
  mode   => '0644',
  source => "/etc/puppetlabs/puppet/ssl/private_keys/${facts['networking']['fqdn']}.pem",
}
file { '/etc/puppetboard/cert.pem':
  ensure => file,
  mode   => '0644',
  source => "/etc/puppetlabs/puppet/ssl/certs/${facts['networking']['fqdn']}.pem",
}

include docker

docker::image { 'ghcr.io/dennispohlmann/puppetboard': }

docker::run { 'puppetboard':
  image   => 'ghcr.io/dennispohlmann/puppetboard',
  volumes => ['/etc/puppetboard:/etc/puppetboard:ro'],
  env     => [
    'PUPPETDB_HOST=puppet', # this must be the certname or DNS_ALT_NAME of the PuppetDB host
    'PUPPETDB_PORT=8081',
    'PUPPETBOARD_PORT=8080',
    'ENABLE_CATALOG=true',
    'PUPPETDB_SSL_VERIFY=false',
    'PUPPETDB_KEY=/etc/puppetboard/key.pem',
    'PUPPETDB_CERT=/etc/puppetboard/cert.pem',
    'SECRET_KEY=XXXXXXXX',
    'DEFAULT_ENVIRONMENT=*',
  ],
  net     => 'host',
}

Within an air gapped environment you want to load all content from your local puppetboard web service. Add: 'OFFLINE_MODE=true', to the env parameter list of the docker::run Puppet type.

We also provide the Dockerfile, so you can build the image yourself:

docker build -t puppetboard .

Manually

Install the package from PyPI into a virtual environment. It brings gunicorn and everything the MCP server needs; the package to import and run is still puppetboard.

python3 -m venv /opt/puppetboard
/opt/puppetboard/bin/pip install puppetboard-dp
PUPPETBOARD_SETTINGS=/etc/puppetboard/settings.py \
  /opt/puppetboard/bin/gunicorn -k asgi -b 127.0.0.1:8080 puppetboard.asgi:app

puppetboard.asgi:app serves the web interface, the JSON API and the MCP server. The plain WSGI app puppetboard.app:app (without MCP) works with any WSGI server. See Deployment setups for a systemd unit and a reverse proxy with authentication.

Configuration

Puppet agents

The default value of usecacheonfailure = true configuration setting for Puppet agents causes Puppet runs to always succeed, event if there are catalog compilation failures f.e. because of a syntax error in your code. This is because in such cases with this setting Puppet will just use a cached working catalog and report the run to PuppetDB as successful. (Although with an error visible in the Puppet run log.)

Therefore, to show the nodes with a catalog compilation as failed in Puppetboard you need to set usecacheonfailure = false in your nodes' puppet.conf.

PuppetDB

Of course you need to configure your Puppet Server to store the Puppet run reports in PuppetDB. If you haven't done that already please follow the PuppetDB documentation about this.

If you run Puppetboard on a different host than PuppetDB then you may want to configure the certificate allow-list for which certificates are allowed to access data from PuppetDB. Please read more about this feature in the PuppetDB documentation here.

App settings

Puppetboard will look for a file pointed at by the PUPPETBOARD_SETTINGS environment variable. The file has to be identical to default_settings.py but should only override the settings you need changed.

If you run PuppetDB and Puppetboard on the same machine the default settings provided will be enough to get you started and you won't need a custom settings file.

Assuming your webserver and PuppetDB machine are not identical you will at least have to change the following settings:

  • PUPPETDB_HOST
  • PUPPETDB_PORT

By default PuppetDB requires SSL to be used when a non-local client wants to connect. Therefore you'll also have to supply the following settings:

  • PUPPETDB_SSL_VERIFY = True
  • PUPPETDB_KEY = /path/to/private/keyfile.pem
  • PUPPETDB_CERT = /path/to/public/keyfile.crt

When using the Puppetboard Docker image, you may also pass Puppetboard it's certificate contents via these environment variables, either as a multiline string or pre-base64 encoded. This can be useful where the certificate is stored in a secrets store i.e. AWS SSM Parameter Store.

PUPPETDB_CERT="-----BEGIN CERTIFICATE-----
...
-----END CERTIFICATE-----"
PUPPETDB_CERT=LS0tLS1CRUdJTiBSU0EgUFJJVkFURSBLRVktLS0tLQouLi4KLS0tLS1FTkQgQ0VSVElGSUNBVEUtLS0tLQ==

For information about how to generate the correct keys please refer to the pypuppetdb documentation. Alternatively it is possible to explicitly specify the protocol to be used setting the PUPPETDB_PROTO variable.

Other settings that might be interesting, in no particular order:

  • SECRET_KEY: set this to a long string, the same for each application replica, and keep it secret. Refer to Flask documentation, section "How to generate good secret keys" for more info. Cannot be an empty string, which is the default.
  • FAVORITE_ENVS: an ordered list of Puppet environment names that will be shown immediately after "All Environments" and before other environments (which are sorted by name) in the dropdown for choosing the environment shown in the top-right of the UI. Environments listed here that do not really exist in your deployment are silently ignored.
  • SHOW_ERROR_AS: friendly or raw. The former makes Puppet run errors in Report and Failures views shown in a modified, (arguably) more user-friendly form. The latter shows them as they are. Defaults to friendly.
  • CODE_PREFIX_TO_REMOVE: what code path that should be shortened in "Friendly errors" to "…" for readability. A regexp. Defaults to /etc/puppetlabs/code/environments(/.*?/modules)?.
  • PUPPETDB_TIMEOUT: Defaults to 20 seconds, but you might need to increase this value. It depends on how big the results are when querying PuppetDB.
  • UNRESPONSIVE_HOURS: The amount of hours since the last check-in after which a node is considered unresponsive.
  • LOGLEVEL: A string representing the loglevel. It defaults to 'info' but can be changed to 'warning' or 'critical' for less verbose logging or 'debug' for more information.
  • ENABLE_QUERY: Defaults to True causing a Query tab to show up in the web interface allowing users to write and execute arbitrary queries against a set of endpoints in PuppetDB. Change this to False to disable this. See ENABLED_QUERY_ENDPOINTS to fine-tune which endpoints are allowed. PQL (Query tab, /api/v1/query, MCP run_pql) can read every PuppetDB entity, catalogs and resource parameters included, also when ENABLE_CATALOG is False. Turn the query off if your catalogs hold secrets.
  • ENABLED_QUERY_ENDPOINTS: If ENABLE_QUERY is True, allow to fine tune the endpoints of PuppetDB APIs that can be queried. It must be a list of strings of PuppetDB endpoints for which the query is enabled. See the QUERY_ENDPOINTS constant in the puppetboard.forms module for a list of the available endpoints.
  • QUERY_PRESETS_FILE: Path to a YAML file with saved queries offered on the Query page, see QUERY_PRESETS.md. Defaults to None (no presets).
  • GRAPH_TYPE: Specify the type of graph to display on fact pages. Default is pie, another good option is donut; the chart library is billboard.js.
  • GRAPH_FACTS: A list of fact names to tell PuppetBoard to generate a pie-chart on the fact page. With some fact values being unique per node, like ipaddress, uuid, and serial number, as well as structured facts it was no longer feasible to generate a graph for everything.
  • INVENTORY_FACTS: A list of tuples that serve as the column header and the fact name to search for to create the inventory page. If a fact is not found for a node then undef is printed.
  • INVENTORY_FACT_TEMPLATES: A mapping between fact name and jinja template to customize display
  • ENABLE_CATALOG: If set to True allows the user to view a node's latest catalog. This includes all managed resources, their file-system locations and their relationships, if available. Also gates catalog resources in the API and MCP server. Defaults to False.
  • REFRESH_RATE: Defaults to 30 the number of seconds to wait until the index page is automatically refreshed.
  • DEFAULT_ENVIRONMENT: Defaults to 'production', as the name suggests, load all information filtered by this environment value, for All Environments use '*'.
  • PAGE_TITLE: Title of the browser window. Defaults to Puppetboard.
  • OVERVIEW_FILTER: An extra PuppetDB query (pypuppetdb QueryBuilder) applied to the nodes on the Overview page. Defaults to None.
  • OFFLINE_MODE: If set to True load static assets (jQuery, Fomantic-UI, DataTables, ...) from the local web server instead of a CDN. Defaults to False.
  • DAILY_REPORTS_CHART_ENABLED: Enable the use of daily chart graphs when looking at dashboard and node view.
  • DAILY_REPORTS_CHART_DAYS: Number of days to show history for on the daily report graphs.
  • DAILY_REPORTS_CHART_MODE: Default layout of the daily report graphs, 'stacked' or 'split'.
  • DISPLAYED_METRICS: Metrics to show when displaying node summary. Example: 'resources.total', 'events.noop'.
  • TABLE_COUNT_SELECTOR: Configure the dropdown to limit number of hosts to show per page.
  • LITTLE_TABLE_COUNT: Default number of reports to show when when looking at a node.
  • NORMAL_TABLE_COUNT: Default number of nodes to show when displaying reports and catalog nodes (container variable REPORTS_COUNT). Defaults to 100.
  • LOCALISE_TIMESTAMP: If set to True then timestamps are shown using your browser's timezone. Otherwise UTC is used. Defaults to True.
  • WITH_EVENT_NUMBERS: If set to True then Overview and Nodes list shows exact number of changed resources in the last report. Otherwise shows only 'some' string if there are resources with given status. Setting this to False gives performance benefits, especially in big Puppet environments (more than few hundreds of nodes). Defaults to True.
  • ENABLE_API: If set to True serves the read-only JSON API under /api/v1/. Defaults to False. See docs/API.md for the endpoints and the API_* settings.
  • ENABLE_MCP: If set to True serves a read-only MCP server under /mcp (ASGI entry point only). Needs MCP_ALLOWED_HOSTS. Defaults to False. See docs/API.md.
  • ENABLE_CLASS: If set to True allows the user to view the number of resource events (number of changed resources in the last report) grouped by class. The resource events are grouped by their status ('failure', 'success', 'noop').
  • CLASS_EVENTS_STATUS_COLUMNS: A mapping between the status of the resource events and the name of the columns of the table to display.
  • CACHE_TYPE: Specifies which type of caching object to use when SCHEDULER_ENABLED is set to True. The cache is used for the classes view (ENABLE_CLASS is set to True) which requires parsing the events of all the latest reports to group them by Puppet class. If the last report is present in the cache, we do not parse the events, which avoids unnecessary processing. If you configure more than one worker, you must use a shared backend (e.g. MemcachedCache) to allow the sharing of the cache between the processes. Indeed, the SimpleCache type does not allow sharing the cache between processes, it uses the process memory to store the cache. Defaults to SimpleCache.
  • CACHE_DEFAULT_TIMEOUT: Cache lifetime in second. Defaults to 3600.
  • SCHEDULER_ENABLED: If set to True then a scheduler instance is created in order to execute scheduled jobs. Defaults to False.
  • SCHEDULER_JOBS: List of the scheduled jobs to trigger within a worker. A job can for example be used to compute a result to be cached. This is the case for the classes view which uses a job to pre-compute at regular intervals the results to be displayed. Each scheduled job must contain the following fields: id, func, trigger, seconds.
  • SCHEDULER_LOCK_BIND_PORT: Specifies an available port that allows a single worker to listen on it. This allows to configure scheduled jobs in a single worker. Defaults to 49100.

Getting Help

For questions or bug reports you can file an issue. Security problems please report privately, see SECURITY.md.

Contributing

Development

Puppetboard relies on the pypuppetdb library to fetch data from PuppetDB and is built with the help of the Flask microframework.

The recommended way to hack on Puppetboard is to:

  1. clone the GitHub repository;
  2. create a virtual environment;
$ python3 -m venv venv
$ . venv/bin/activate # run `deactivate` to exit the virtual environment
  1. install the package and the test requirements.
$ pip install -r requirements-test.txt -e .

You can then run the tests with:

pytest --cov=. --cov-report=xml --strict-markers --mypy puppetboard test
pylint --errors-only puppetboard test

And you can run the app by executing:

$ export PUPPETBOARD_SETTINGS=$PWD/settings.py # see settings.py.sample
$ flask --app puppetboard.app run

Host and port can be set with --host and --port. Every change comes with matching unit tests.

License

This project is licensed under the Apache v2.0 License. As such, your contributions, once accepted, are automatically covered by this license. Bundled third-party files and their licenses are listed in THIRD_PARTY.md.

Commit messages

Write decent commit messages. Don't use swear words and refrain from uninformative commit messages as 'fixed typo'.

The preferred format of a commit message:

Fixed a typo in the Nodes section.

If needed, elaborate further on this commit. Feel free to write a
complete blog post here if that helps us understand what this is
all about.

Fixes #4 and resolves #2.

If you'd like a more elaborate guide on how to write and format your commit messages have a look at this post by Tim Pope.

Build a release

please see: RELEASE.md

More Screenshots

  • Overview / Index / Homepage

Overview / Index / Homepage

  • Nodes view, all active nodes

Nodes view, all active nodes

  • Single node page / overview

Single node page / overview

  • Report view

Report view

  • Failures view, the errors of all failing nodes

Failures view

  • Facts view

Facts view

  • Single fact, with graphs

Single fact, with graphs

  • All nodes that have this fact with that value

All nodes that have this fact with that value

  • Query view - results as table

Query view

  • Query view - results as JSON

Query view

  • Metrics view

Metrics view

  • Single metric

Single metric

  • Inventory view

Inventory view

  • Classes view, group the resource events of the last reports by Puppet class

Classes view

  • Class view, list the nodes with almost one resource event for a given class

Class view

  • Radiator view, for a wall display

Radiator view

Legal

The app code is licensed under the Apache License, Version 2.0.

The favicon has been created based on the icon created by ShellyCoen under the Creative Commons Attribution-ShareAlike 4.0 International License, downloaded from the voxpupuli/logos, resized and centered.

Metadata

Release files for puppetboard-dp 8.2.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for puppetboard-dp 8.2.1
File Size Uploaded Provenance Actions
puppetboard_dp-8.2.1.tar.gz 3.8 MB With provenance Details for puppetboard_dp-8.2.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for puppetboard-dp 8.2.1
File Interpreter ABI Platform Actions
puppetboard_dp-8.2.1-py3-none-any.whl Python 3 none any Details for puppetboard_dp-8.2.1-py3-none-any.whl

Total release size: 7.6 MB

Release files / puppetboard_dp-8.2.1.tar.gz

Download URL puppetboard_dp-8.2.1.tar.gz
Size 3.8 MB
Tags Source
SHA-256 checksum
How to use checksums
ee5694a0bb6629d389957f2708641c8e0eb299c873760bd0aebefc36625456af
BLAKE2b-256 checksum
How to use checksums
501a4fe016db34609c8d39592a99acdb6116038e55b124a4b18364da797253eb
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 19, 2026.

Transparency log

Release files / puppetboard_dp-8.2.1-py3-none-any.whl

Download URL puppetboard_dp-8.2.1-py3-none-any.whl
Size 3.8 MB
Tags Python 3
SHA-256 checksum
How to use checksums
6d75c67ab3f45a91d86e0c518b3f39ec313c051edcc6fcbaf7ee429a2969611f
BLAKE2b-256 checksum
How to use checksums
046af03f08e3e94ac106685c6a07e189e00416b1a43326878930640635135ae7
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 19, 2026.

Transparency log

Security for puppetboard-dp 8.2.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for puppetboard_dp-8.2.1.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 19, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page