consul/website/source/docs/agent/acl-system.html.md

12 KiB

layout page_title sidebar_current description
docs ACL System docs-agent-acl-system Consul provides an optional Access Control List (ACL) system which can be used to control access to data and APIs. The ACL system is a Capability-based system that relies on tokens which can have fine grained rules applied to them. It is very similar to AWS IAM in many ways.

-> 1.4.0 and later: This guide only applies in Consul versions 1.4.0 and later. The documentation for the legacy ACL system is here

ACL System

Consul provides an optional Access Control List (ACL) system which can be used to control access to data and APIs. The ACL is Capability-based, relying on tokens which are associated with policies to determine which fine grained rules can be applied. Consul's capability based ACL system is very similar to the design of AWS IAM.

ACL System Overview

The ACL system is designed to be easy to use and fast to enforce while providing administrative insight. At the highest level, there are two major components to the ACL system:

  • ACL Policies - Policies allow the grouping of a set of rules into a logical unit that can be reused and linked with many tokens.

  • ACL Tokens - Requests to Consul are authorized by using bearer token. Each ACL token has a public Accessor ID which is used to name a token, and a Secret ID which is used as the bearer token used to make requests to Consul.

ACL tokens and policies are managed by Consul operators via Consul's ACL API, ACL CLI, or systems like HashiCorp's Vault.

ACL Policies

An ACL policy is a named set of rules and is composed of the following elements:

  • ID - The policies auto-generated public identifier.
  • Name - A unique meaningful name for the policy.
  • Rules - Set of rules granting or denying permissions. See the Rule Specification documentation for more details.
  • Datacenters - A list of datacenters the policy is valid within.

Builtin Policies

  • Global Management - Grants unrestricted privileges to any token that uses it. When created it will be named global-management and will be assigned the reserved ID of 00000000-0000-0000-0000-000000000001. This policy can be renamed but modification of anything else including the rule set and datacenter scoping will be prevented by Consul.

ACL Tokens

ACL tokens are used to determine if the caller is authorized to perform an action. An ACL token is composed of the following elements:

  • Accessor ID - The token's public identifier.
  • Secret ID -The bearer token used when making requests to Consul.
  • Description - A human readable description of the token. (Optional)
  • Policy Set - The list of policies that are applicable for the token.
  • Locality - Indicates whether the token should be local to the datacenter it was created within or created in the primary datacenter and globally replicated.

Builtin Tokens

During cluster bootstrapping when ACLs are enabled both the special anonymous and the master token will be injected.

  • Anonymous Token - The anonymous token is used when a request is made to Consul without specifying a bearer token. The anonymous token's description and policies may be updated but Consul will prevent this tokens deletion. When created, it will be assigned 00000000-0000-0000-0000-000000000002 for its Accessor ID and anonymous for its Secret ID.

  • Master Token - When a master token is present within the Consul configuration, it is created and will be linked With the builtin Global Management policy giving it unrestricted privileges. The master token is created with the Secret ID set to the value of the configuration entry.

Authorization

The token Secret ID is passed along with each RPC request to the servers. Consul's HTTP endpoints can accept tokens via the token query string parameter, the X-Consul-Token request header, or an RFC6750 authorization bearer token. Consul's CLI commands can accept tokens via the token argument, or the CONSUL_HTTP_TOKEN environment variable.

If no token is provided for an HTTP request then Consul will use the default ACL token if it has been configured. If no default ACL token was configured then the anonymous token will be used.

ACL Rules and Scope

The rules from all policies linked with a token are combined to form that token's effective rule set. Policy rules can be defined in either a whitelist or blacklist mode depending on the configuration of acl_default_policy. If the default policy is to "deny" access to all resources, then policy rules can be set to whitelist access to specific resources. Conversely, if the default policy is “allow” then policy rules can be used to explicitly deny access to resources.

The following table summarizes the ACL resources that are available for constructing rules:

Resource Scope
acl Operations for managing the ACL system ACL API
agent Utility operations in the Agent API, other than service and check registration
event Listing and firing events in the Event API
key Key/value store operations in the KV Store API
keyring Keyring operations in the Keyring API
node Node-level catalog operations in the Catalog API, Health API, Prepared Query API, Network Coordinate API, and Agent API
operator Cluster-level operations in the Operator API, other than the Keyring API
query Prepared query operations in the Prepared Query API
service Service-level catalog operations in the Catalog API, Health API, Prepared Query API, and Agent API
session Session operations in the Session API

Since Consul snapshots actually contain ACL tokens, the Snapshot API requires a token with "write" privileges for the ACL system.

The following resources are not covered by ACL policies:

  1. The Status API is used by servers when bootstrapping and exposes basic IP and port information about the servers, and does not allow modification of any state.

  2. The datacenter listing operation of the Catalog API similarly exposes the names of known Consul datacenters, and does not allow modification of any state.

  3. The connect CA roots endpoint exposes just the public TLS certificate which other systems can use to verify the TLS connection with Consul.

Constructing rules from these policies is covered in detail in the Rule Specification section below.

Configuring ACLs

ACLs are configured using several different configuration options. These are marked as to whether they are set on servers, clients, or both.

Configuration Option Servers Clients Purpose
acl.enabled REQUIRED REQUIRED Controls whether ACLs are enabled
acl.default_policy OPTIONAL N/A Determines whitelist or blacklist mode
acl.down_policy OPTIONAL OPTIONAL Determines what to do when the remote token or policy resolution fails
acl.policy_ttl OPTIONAL OPTIONAL Determines time-to-live for cached ACL Policies
acl.token_ttl OPTIONAL OPTIONAL Determines time-to-live for cached ACL Tokens

A number of special tokens can also be configured which allow for bootstrapping the ACL system, or accessing Consul in special situations:

Special Token Servers Clients Purpose
acl.tokens.agent_master OPTIONAL OPTIONAL Special token that can be used to access Agent API when remote bearer token resolution fails; used for setting up the cluster such as doing initial join operations, see the ACL Agent Master Token section for more details
acl.tokens.agent OPTIONAL OPTIONAL Special token that is used for an agent's internal operations, see the ACL Agent Token section for more details
acl.tokens.master OPTIONAL N/A Special token used to bootstrap the ACL system, see the Bootstrapping ACLs section for more details
acl.tokens.default OPTIONAL OPTIONAL Default token to use for client requests where no token is supplied; this is often configured with read-only access to services to enable DNS service discovery on agents

All of these tokens except the master token can all be introduced or updated via the /v1/agent/token API.

ACL Agent Master Token

Since the acl.tokens.agent_master is designed to be used when the Consul servers are not available, its policy is managed locally on the agent and does not need to have a token defined on the Consul servers via the ACL API. Once set, it implicitly has the following policy associated with it

agent "<node name of agent>" {
  policy = "write"
}
node_prefix "" {
  policy = "read"
}

ACL Agent Token

The acl.tokens.agent is a special token that is used for an agent's internal operations. It isn't used directly for any user-initiated operations like the acl.tokens.default, though if the acl.tokens.agent_token isn't configured the acl.tokens.default will be used. The ACL agent token is used for the following operations by the agent:

  1. Updating the agent's node entry using the Catalog API, including updating its node metadata, tagged addresses, and network coordinates
  2. Performing anti-entropy syncing, in particular reading the node metadata and services registered with the catalog
  3. Reading and writing the special _rexec section of the KV store when executing consul exec commands

Here's an example policy sufficient to accomplish the above for a node called mynode:

node "mynode" {
  policy = "write"
}
service_prefix "" {
  policy = "read"
}
key_prefix "_rexec" {
  policy = "write"
}

The service_prefix policy needs read access for any services that can be registered on the agent. If remote exec is disabled, the default, then the key_prefix policy can be omitted.

Next Steps

Setup ACLs with the Boostrapping guide or continue reading about ACL rules.