Apply an Astro Private Cloud platform Helm configuration change

Astro Private Cloud uses Helm to install and manage some platform-level settings that apply to all users and Deployments. The Astronomer Helm chart includes configurations for areas such as:

Identity provider integrations
Registry backends
Resource allocation limits

You can apply any platform customizations to your cluster using a YAML configuration. values.yaml contains settings for both the Astronomer Helm chart, as well as Helm charts for system components like Elasticsearch and nginx. Using Helm allows you to keep all of your configurations in a single file that you can version and store securely.

Use this document to learn how to retrieve your existing Helm configuration, modify configurations, and apply your changes to your cluster.

Before you edit values.yaml, read Configure Astro Private Cloud to decide whether a setting belongs in platform Helm values, a cluster override, a workspace override, or a deployment override.

Retrieve your current cluster configuration

The best way to add or modify configurations is to start with your existing values.yaml file. To retrieve the current values.yaml file of an existing APC cluster, run the following command:

$ helm get values -o yaml <your-installation-release-name> -n <your-installation-namespace> > values.yaml

This command overwrites any file called values.yaml in your current directory.

Alternatively, your team might use version management to store your existing Helm configuration. In this case, retrieve, update, and store your configuration file according to your team’s workflows. If your configuration file includes secrets, ensure you encrypt it before storing it in version control.

Update your configurations

Create a copy of your values.yaml file so that you can compare your existing configuration to your new configuration.
In your copied file, update the values for the configurations you want to change. To update a configuration you haven’t already specified, copy the corresponding default values from the relevant default Helm chart into your values.yaml file and then modify the value.

When you have finished updating the configuration, ensure that the configurations have the same relative order and indentation as they do in the default configuration file. If they don’t, your changes might not be properly applied. The name of the Helm charts you’re modifying should be the first items in the file, such as in the following example:

1 global:
2   <your-global-configuration>
3 
4 astronomer:
5   <your-astronomer-configuration>
6 
7 alertmanager:
8   <your-alertmanager-configuration>
9 
10 nginx:
11   <your-nginx-configuration>

A number of configurations, including identity provider settings and platform-wide deployment defaults, live in astronomer.houston.config. Astro Private Cloud (APC) 2.x applies platform config first, then cluster, workspace, and Deployment overrides for supported deployments.* settings. Use Configure Astro Private Cloud to decide which layer to update before you run a Helm upgrade.

Push changes to your cluster

Copy your platform namespace and release name. These are both likely to be astronomer. Find your platform release name in your list of active namespaces by running the following command:

$ kubectl get ns

To locate your platform release name, run:

$ helm ls -n <your-installation-namespace>

Save your updated values.yaml file and run the following command to apply it as a Helm upgrade:

$ helm upgrade <your-installation-release-name> astronomer/astronomer -f <your-updated-config-yaml-file> -n <your-installation-namespace> --set astronomer.houston.upgradeDeployments.enabled=false

Setting astronomer.houston.upgradeDeployments.enabled=false ensures that Apache Airflow components in Deployments don’t restart during your upgrade.

Run the following command to confirm that your configuration was applied:

$ helm get values <your-installation-release-name> -n <your-installation-namespace>

1	global:
2	<your-global-configuration>
3
4	astronomer:
5	<your-astronomer-configuration>
6
7	alertmanager:
8	<your-alertmanager-configuration>
9
10	nginx:
11	<your-nginx-configuration>