Each environment keeps a list of the resources it manages, by their key in the file
(resources.db, and a service’s own resources as services.<name>.<kind>). ezgh deploy status
and the canvas show them with their status.
| Status | Means |
|---|---|
creating, changing |
A deployment is creating or changing it |
managed |
The environment manages it |
createFailed |
Creating it failed; the next deployment tries again |
detaching, detached |
It was removed from the file; Deploy no longer changes it |
missing |
Its product no longer has it |
deleting, deleted |
It’s being deleted, or was, on request |
forgotten |
Deploy stopped tracking it; the resource itself is untouched |
On product pages
Every product page for a managed resource says so: “Managed by Deploy stack shop / production.
Changes made here are reverted by its next deployment.” It links to the stack. The file wins for
each field it sets, so change those in the file. Fields the file doesn’t set can be changed on
the product’s page and are kept.
Resources whose products take tags carry ezgh-deploy-stack, ezgh-deploy-environment and
ezgh-deploy-entry. The environment’s list is what counts: bots, buckets, alert rules, listeners
and firewall rules carry no tags.
Detached
Removing a resource from the file detaches it at the next deployment. Deploy stops managing it, removes its tags, and leaves it and its data as they are. A detached resource shows on the canvas until it’s attached again or dismissed (forgotten).
Missing
A database, cache, bucket, queue or secret that its product no longer has is missing, and is
never recreated by itself: plans that touch it are refused until you forget or attach it, so a
lost database is never replaced by an empty one without you knowing. A missing firewall, target
group, listener or alert rule is recreated by the next deployment, and shown in its plan.
Attach
ezgh deploy resources attach db dbi_0123456789abhas the environment manage an existing resource under a key in the file, instead of creating one. You need the product’s read and update rights on that resource, and it can’t already be managed by another environment. Its next deployment’s plan shows what changes to match the file.
Forget
ezgh deploy resources forget dbstops managing a resource and leaves it as it is. The CLI names the stack, environment, kind and
ID, and asks you to type <environment>/<key>; --yes skips the question. Forgetting a missing
resource lets the next plan create a new one.
Delete
ezgh deploy resources delete db --confirm production/dbdeletes a managed resource, as its own deployment. It’s the only way Deploy deletes one. The CLI
names the stack, environment, kind and ID, and needs <environment>/<key> typed or passed with
--confirm. You need the product’s own delete right on the resource, as well as
deploy.resources.delete.
| Action | Permission |
|---|---|
| List | deploy.resources.list |
| Attach | deploy.resources.attach, and the product’s read and update rights on the resource |
| Forget | deploy.resources.forget |
| Delete | deploy.resources.delete, and the product’s delete right on the resource |