Databases and caches
Canopy provisions database and cache service instances in an environment. The supported engines are PostgreSQL, Redis, and an externally managed service. Provisioning records the requested state; the Grove that owns the environment applies it on a later poll, so the resource may not be ready when the command returns.
Create and inspect a service
Section titled “Create and inspect a service”Create a PostgreSQL instance:
canopy services create db --engine postgres --env <environment-id>Create Redis with a storage size:
canopy services create cache --engine redis --size-gb 5 --env <environment-id>List or inspect service instances:
canopy services list --env <environment-id>canopy services status <service-instance-id> --env <environment-id>The CLI also accepts --provider when you need to select a supported provider explicitly. PostgreSQL resource settings can be set with --cpu, --memory, and --memory-limit. See the CLI reference for the available flags.
Connect an application
Section titled “Connect an application”Creating a service does not automatically add a connection URL to an application’s variables. Configure a service binding in the workload’s .canopy.yml services section, or set the required connection variables yourself. The binding’s names: conventional mode requests conventional variables such as DATABASE_URL for PostgreSQL and REDIS_URL for Redis. See Environments and services for binding configuration and variable management.
Treat connection strings and credentials as secrets. Do not put their values in a committed manifest or documentation example. Use Canopy’s secret variable handling for credentials you set yourself.
Delete a service
Section titled “Delete a service”canopy services destroy <service-instance-id> --env <environment-id>Destroying a service schedules its deletion by the owning Grove. Back up data you need before deleting the service.
Backups and restore
Section titled “Backups and restore”Canopy exposes service backup and restore operations, but availability depends on the service’s backup target configuration. A backup request records work for the Grove; it does not mean a backup has completed. Do not assume that every database has automatic scheduled backups or a fixed retention period.
List backups and restore a successful backup to a new service instance:
canopy services backup <service-instance-id> --env <environment-id>canopy services backups <service-instance-id> --env <environment-id>canopy services restore <service-instance-id> \ --backup <backup-id> \ --name db-restored \ --env <environment-id>The restored instance is provisioned asynchronously. Verify it before changing your application to use it. Point-in-time restore is available only when the source backup and its WAL archive support the requested time; pass --target-time <ISO-8601> when applicable. Ask your Canopy administrator or support contact whether backups are configured for your environment and what retention applies.
Test databases
Section titled “Test databases”Canopy does not provide a customer-facing command or a separately provisioned environment test database for this purpose. TESTING_DATABASE_URL and TEST_DATABASE_URL are configuration names used by Canopy’s own API test tooling; they are not customer environment variables automatically injected by database provisioning.