Docker Authorization

Docker’s out-of-the-box authorization model is all or nothing. But many users require finer-grained access control and Docker’s plugin infrastructure allows us to do so.

This is an excellent opportunity to see how to policy enable an existing service.


This tutorial helps you get started with OPA and introduces you to core concepts in OPA, including Rego the language used to define policies.

Policy enabling an application decouples the policy implementation from the business logic so that administrators can define policy without changing the application while still keeping up with the size, complexity, and dynamic nature of modern applications.

Although there are a multitude of desirable access control policies, for demonstration purposes, we want to prevent the following:

  • Containers with insecure configurations.
  • Users modifying the system without sufficient read+write access.

This tutorial illustrates two key concepts:

  1. OPA policy definition is decoupled from the implementation of the service (in this case Docker). The administrator is empowered to define and manage policies without requiring changes to any of the apps.
  2. Both the data relevant to policy and the policy definitions themselves can change rapidly.

Once you finish this tutorial, you will be familiar with:

  • Running OPA as a server/daemon.
  • Loading policy definitions and data via the REST APIs.
  • Querying data via the REST APIs.
  • The basics of Rego, OPA’s purpose-built policy language.


This tutorial requires:

  • Docker Engine 1.11 or newer
  • root or sudo access

The tutorial has been tested on the following platforms:

  • Ubuntu 16.04 (64-bit)

If you are using a different distro, OS, or architecture, the steps will be the same. However, there may be slight differences in the commands you need to run.


1. Download the latest version of OPA.

$ curl -L > opa
$ chmod u+x opa

2. Run OPA in server mode with debug logging enabled.

$ ./opa run --server --log-level debug

OPA will run until it receives a signal to stop. Open another terminal to continue with the rest of the tutorial.

3. Download the open-policy-agent/opa-docker-authz executable.

$ curl -L > opa-docker-authz
$ chmod u+x opa-docker-authz

The open-policy-agent/opa-docker-authz repository hosts a small Docker Authorization Plugin. Docker's authorization plugin system allows an external process to receive all requests sent to the Docker daemon. The authorization plugin replies, instructing the Docker daemon to allow or reject the request.

4. Create an empty policy definition that will allow all requests.

$ cat >example.rego <<EOF
package opa.example

allow_request = true

This policy definition is about simple as it can be. It includes a single rule named allow_request that is defined to always be true. Once all of the components are running, we will come back and extend this policy.

5. Run the opa-docker-authz plugin and then open another terminal.

$ sudo ./opa-docker-authz -policy-file=example.rego

This step requires sudo access because the Docker plugin framework will attempt to update the Docker daemon configuration. If you run without sudo you may encounter a permission error.

6. Reconfigure Docker.

Docker must include the following command-line argument:


On Ubuntu 16.04 with systemd, this can be done as follows (requires root):

$ sudo mkdir -p /etc/systemd/system/docker.service.d
$ sudo tee -a /etc/systemd/system/docker.service.d/override.conf > /dev/null <<EOF
ExecStart=/usr/bin/docker daemon -H fd:// --authorization-plugin=opa-docker-authz
$ sudo systemctl daemon-reload
$ sudo service docker restart

If you are using a different Linux distribution or you are not running systemd, the process will be slightly different.

7. Run a simple Docker command to make sure everything is still working.

$ docker ps

If everything is setup correctly, the command should exit successfully. You can expect to see log messages from OPA and the plugin.

8. Test that the policy definition is working.

Let’s modify our policy to deny all requests:

$ cat >example.rego <<EOF
package opa.example

allow_request = true { false }

In OPA, rules defines the content of documents (for example, objects, arrays, strings, booleans, and so on).

In steps above, we created a rule named allow_request that defines a document that is the boolean value true. When all of the expressions on the body of the rule evaluate to true, we say the document is defined. If any of the expressions evaluate to false, we say the document is undefined. In this case, the document will always be undefined because the body of the rule is false.

$ docker ps

The output should be:

Error response from daemon: authorization denied by plugin opa-docker-authz: request rejected by administrative policy

To learn more about how rules define the content of documents, see: How Does OPA Work?

With this policy in place, users will not be able to run any Docker commands. Go ahead and try other commands such as docker run or docker pull. They will all be rejected.

Now let's change the policy so that it's a bit more useful.

9. Update the policy to reject requests with the unconfined seccomp profile:

$ cat >example.rego <<EOF
package opa.example

seccomp_unconfined {
    # This expression asserts that the string on the right-hand side is equal
    # to an element in the array SecurityOpt referenced on the left-hand side.
    input.Body.HostConfig.SecurityOpt[_] = "seccomp:unconfined"

allow_request {
    not seccomp_unconfined

The opa-docker-authz plugin is watching the policy definition file for changes. Each time we change the file, the plugin reads the file and sends it to OPA. To manually send the policy to OPA, you can use the following API:

$ curl -X PUT --data-binary @example.rego http://localhost:8181/v1/policies/example_policy

This API is idempotent so sending the policy multiple times is fine. Go ahead and try it yourself.

10. Test the policy is working by running a simple container:

$ docker run hello-world

Now try running the same container but disable seccomp (which should be prevented by the policy):

$ docker run --security-opt seccomp:unconfined hello-world

When Docker processes the run command, it contacts the plugin to see if the request should be allowed. The plugin takes the request and executes a query against OPA using the request as input data to the query. The same API call that the plugin makes can be executed using curl:

$ curl -i -d '{"input": {"Body":{"HostConfig":{"SecurityOpt":["seccomp:unconfined"]}}}}' \
    -H 'Content-Type: application/json' \

Because the document generated by the allow_request rule is undefined in this case, the response object does not include a value for the document.

You can re-run the same query with the default seccomp profile and see that it succeeds:

$ curl -i -d '{"input": {"Body":{"HostConfig":{"SecurityOpt":["seccomp:default"]}}}}' \
    -H 'Content-Type: application/json' \

Congratulations! You have successfully prevented containers from running without seccomp!

So far, the policy has been defined in terms of input data from the plugin. In many cases, it's necessary to write policies against multiple data sources.

The rest of the tutorial shows how you can grant fine grained access to specific clients. To do so, we will insert fake user data into OPA to simulate an authentication system.

11. Identify the user in Docker requests.

Back up your existing Docker configuration, just in case. You can replace your original configuration after you are done with the tutorial.

$ mkdir -p ~/.docker
$ cp ~/.docker/config.json ~/.docker/config.json~

To identify the user, include an HTTP header in all of the requests sent to the Docker daemon:

$ cat >~/.docker/config.json <<EOF
    "HttpHeaders": {
        "Authz-User": "bob"

Docker does not currently provide a way to authenticate clients. But in Docker 1.12, clients can be authenticated using TLS and there are plans to include other means of authentication. For the purpose of this tutorial, we assume that an authentication system is place.

12. Add user data directly to OPA.

$ cat >users.json <<EOF
        "op": "add",
        "path": "/",
        "value": {
            "alice": {
                "readOnly": false
            "bob": {
                "readOnly": true
$ curl -X PATCH -d @users.json http://localhost:8181/v1/data/users -H "Content-Type: application/json"

This data represents information about users that could either come from an external system or be included in policy definitions.

To see that the user data has been added, we can query the Data API. This shows the properties associated with the user "alice":

$ curl http://localhost:8181/v1/data/users/alice

13. Update the policy to include basic user access controls.

$ cat >example.rego <<EOF
package opa.example

import data.users

allow_request {

# valid_user_role defines a document that is the boolean value true if this is
# a write request and the user is allowed to perform writes.
valid_user_role {
    user_id = input.Headers["Authz-User"]
    user = users[user_id]
    user.readOnly = false

# valid_user_role is defined again here to handle read requests. When a rule
# like this is defined multiple times, the rule definition must ensure that
# only one instance evaluates successfully in a given query. If multiple
# instances evaluated successfully, it indicates a conflict.
valid_user_role {
    user_id = input.Headers["Authz-User"]
    user = users[user_id]
    input.Method = "GET"
    user.readOnly = true

In the new policy, the valid_user_role rules reference the "users" document created in the previous step.

14. Attempt to run a container.

Because the configured user is "bob", the request is rejected:

$ docker run hello-world

15. Change the user to "alice" and re-run the container.

$ cat > ~/.docker/config.json <<EOF
    "HttpHeaders": {
        "Authz-User": "alice"

Because the configured user is "alice", the request will succeed:

$ docker run hello-world

16. Restore your original configuration.

See: “Reconfigure Docker.” and “Identify the user in Docker requests.”.

That's it!

results matching ""

    No results matching ""