devops-config-mgmt

name	devops-config-mgmt
description	Routes configuration management requests to the correct technology agent. Compares Chef, Puppet, SaltStack, and Ansible for server configuration and compliance. WHEN: "configuration management", "config management comparison", "Chef vs Puppet", "Chef vs Ansible", "server configuration", "compliance automation", "desired state configuration", "convergence".
license	MIT
metadata	{"version":"1.0.0","author":"christopher huffman"}

Configuration Management Router

You are a routing agent for configuration management technologies. You determine which technology best matches the user's question and delegate to the appropriate specialist.

Decision Matrix

Signal	Route To
Chef, cookbook, recipe, knife, Chef Infra, InSpec, Habitat	chef/SKILL.md
Puppet, manifest, module, Facter, Hiera, PuppetDB, Bolt	puppet/SKILL.md
SaltStack, Salt, minion, grain, pillar, state file, salt-master	saltstack/SKILL.md
Ansible, playbook, role, inventory, AWX	ansible/SKILL.md
Config management comparison, "which tool"	Handle directly (below)

Comparison Matrix

Dimension	Chef	Puppet	SaltStack	Ansible
Language	Ruby DSL	Puppet DSL	YAML + Jinja2	YAML + Jinja2
Architecture	Client-server (agent)	Client-server (agent)	Client-server (agent) or agentless	Agentless (SSH)
Model	Imperative (convergent)	Declarative	Declarative + imperative	Procedural (idempotent)
Agent	chef-client	puppet-agent	salt-minion	None
Pull/Push	Pull (agent polls server)	Pull (agent polls server)	Push or pull	Push (SSH)
Communication	HTTPS (client → server)	HTTPS (agent → master)	ZeroMQ or SSH	SSH / WinRM
Scalability	Good (10K+ nodes)	Good (10K+ nodes)	Excellent (10K+ nodes)	Moderate (SSH limits)
Learning curve	High (Ruby required)	Medium (Puppet DSL)	Medium (YAML + Jinja2)	Low (YAML)
Community	Shrinking	Large, enterprise	Growing	Largest
License	Apache 2.0	Apache 2.0	Apache 2.0	GPL v3

When to Choose

Scenario	Recommended	Why
Greenfield, simple needs	Ansible	Lowest barrier, agentless, huge module ecosystem
Large fleet (10K+ servers)	SaltStack or Puppet	Agent-based scales better, faster execution
Enterprise compliance	Chef (InSpec) or Puppet	Mature compliance frameworks
Windows-heavy environment	Puppet or Ansible	Strong Windows support, DSC integration
Network devices	Ansible	Best network module ecosystem
Cloud-native / containers	Ansible or SaltStack	Better cloud integration, less agent overhead
Existing Ruby team	Chef	Ruby DSL feels natural

Configuration Management Concepts

Load references/concepts.md for foundational CM patterns.

Convergence

All CM tools aim for convergence — bringing a system from its current state to the desired state:

1. Detect current state (package installed? file content? service running?)
2. Compare current vs desired
3. Remediate if different (install, update, restart)
4. Report what changed

Agent vs Agentless

Aspect	Agent-Based (Chef, Puppet, Salt)	Agentless (Ansible)
Continuous enforcement	Agent runs periodically (every 30 min)	Only when playbook runs
Speed at scale	Faster (local execution)	Slower (SSH per host)
Bootstrap	Must install agent first	SSH access sufficient
Firewall	Agent initiates outbound (simpler)	Control node needs inbound SSH
Overhead	Agent process on every node	No overhead on managed nodes

Reference Files

• references/concepts.md — Configuration management theory (convergence, idempotency, desired state, compliance as code, drift remediation)