planetary-computer-apis/deployment

Deployment

This folder contains the code to run a deployment of the Planetary Computer APIs in Azure. It includes a few steps that use Terraform and Helm.

Deploy script

The logic for the deployment workflow is encapsulated in the bin/deploy script. You can run from the top level of the repository via scripts/cideploy by CI. You can also drop into a console to manually run steps for a dev deployment via

scripts/console --deploy

To have access to the remote backend terraform state, the identity (App Registration in CI, or local corp credential if local) will need to have the Storage Blob Data Owner role on the pctesttfstate storage account.

Manual resources

Deployment secrets Key Vault

Deployment requires access to a Key Vault with some deployment secrets. If you are making you own stack, you'll need to create your own key vault and set the appropriate variable in the terraform. See the keyvault.tf resource for the list of secrets required, and to make sure these docs are not out of date.

The service principle here needs to have the Storage Account Data Contributor role assigned for each container made available through the sas-token and sign endpoints.

Dev deployment

To deploy your own stack, terraform will run in the deployment/terraform/dev stack. Run scripts/deploy --dev to do this. You'll need the following environment variables set according with the credential information for a service principal that has permissions to deploy to your subscription:

• subscriptionId
• tenantId
• servicePrincipalId
• servicePrincipalKey

If you want to deploy local code changes, instead of the published images, be sure to set the following environment variables with the correct public Azure Container Registry repo where you published your local images:

• ACR_STAC_REPO
• ACR_TILER_REPO
• IMAGE_TAG

Note: Remember to bring down your resources after testing with terraform destroy!

Loading configuration data

Configuration data is stored in Azure Storage Tables. Use the pcapis command line interface that is installed with the pccommon package to load data. For example:

>  az login # Use an account that has "Storage Table Data Contributor" on the account
>  pcapis load -t collection --account pctapissatyasa  --table collectionconfig --file pccommon/tests/data-files/collection_config.json

To dump a single collection config, use:

>  pcapis dump -t collection  --account pctapissatyasa  --table collectionconfig --id naip

For container configs, you must also specify the container account name used as the Partition Key:

>  pcapis dump -t collection --account pctapissatyasa  --table containerconfig --id naip --container-account naipeuwest

Using the load command on a single dump file for either config will update the single row.