How to Collect Kubernetes Resource Data

The Kubernetes collector is configured within the Fix Inventory Worker configuration via the config command in Fix Inventory Shell.

Prerequisites

This guide assumes that you have already installed Fix Inventory.

Directions

1. Enable the Kubernetes Collector

1. Open the Fix Inventory Worker configuration via the config command in Fix Inventory Shell:

   > config edit fix.worker

2. Add k8s to the list of collectors by modifying the configuration as follows:

   Fix Inventory Worker configuration

   fixworker:
     ...
     # List of collectors to run
     collector:
       - 'k8s'
       ...
   ...

2. Authenticate with Kubernetes

You can authenticate with Kubernetes via kubeconfig files, manual configuration, or both.

kubeconfig Files

The easiest way to configure access to Kubernetes is via kubeconfig files.

1. Open the Fix Inventory Worker configuration via the config command in Fix Inventory Shell:

   > config edit fix.worker

2. Add the content of kubeconfig file(s) to the fixworker section as follows:

   Fix Inventory Worker configuration

   fixworker:
     ...
     write_files_to_home_dir:
       - path: ~/.kube/config_1
         content: |
           apiVersion: v1
           clusters:
           - cluster:
               certificate-authority-data: <ca_data>
               server: https://k8s.example.com
             name: example-cluster
           contexts:
           - context:
               cluster: example-cluster
               user: k8s-admin
             name: context1
           current-context: context1
           kind: Config
           preferences: {}
           users:
           - name: k8s-admin
             user:
               token: <token>
       - path: ~/.kube/config_2
         content: ...
     ...
   ...
   note

   If you do not wish to save the contents of your kubeconfig file(s) to Fix Inventory's database, you can alternatively mount the directory containing your kubeconfig file(s) to the fixworker container.

   info

   Since Fix Inventory is running on your local machine, it can access the file(s) directly. Move or copy your kubeconfig file(s) to the ~/.kube directory.

3. Modify the k8s section of the configuration as follows, defining path and contexts for each file:

   Fix Inventory Worker configuration

   fixworker:
     ...
   ...
   k8s:
     config_files:
       - path: "/home/fixinventory/.kube/config_1"
         all_contexts: false
         contexts: ["context1", "context2"]
       - path: "/home/fixinventory/.kube/config_2"
         all_contexts: true
   note

   The above example assumes that your kubeconfig file(s) are named config_1, config_2, etc.

   info

   If a single kubeconfig file holds multiple contexts, it is possible to restrict the contexts to be used by defining them explicitly. Setting all_contexts to true will not filter, resulting in taking all found contexts.

Manual Configuration

Instead of exposing a kubeconfig file to Fix Inventory Worker, you can alternatively supply credentials manually.

The required values can be found in the kubeconfig file:

Option	kubeconfig Property
server	clusters.cluster.server
token	users.user.token
certificate_authority_data	clusters.cluster.certificate-authority-data (only required if the server is using a self-signed certificate)

1. Open the Fix Inventory Worker configuration via the config command in Fix Inventory Shell:

   > config edit fix.worker

2. Modify the k8s section of the configuration as follows:

   Fix Inventory Worker configuration

   fixworker:
     ...
   ...
   k8s:
     configs:
       - name: 'dev'
         certificate_authority_data: 'xxx'
         server: 'https://k8s-cluster-server.example.com'
         token: 'token'
   info

   Multiple k8s clusters can be defined by adding multiple sets of values.

3. Trigger Resource Collection

By default, Fix Inventory performs resource collection each hour. To immediately trigger a collect run, use the workflow run command in Fix Inventory Shell:

> workflow run collect

Once the collect run completes, you can view a summary of collected Kubernetes resources using the following search:

> search is(kubernetes_resource) | count kind