import requests
url = "https://platform.ankra.app/api/v1/clusters/resources/roll-to"
payload = {
"cluster_id": "<string>",
"version_id": "<string>"
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({cluster_id: '<string>', version_id: '<string>'})
};
fetch('https://platform.ankra.app/api/v1/clusters/resources/roll-to', options)
.then(res => res.json())
.then(res => console.log(res))
.catch(err => console.error(err));curl --request POST \
--url https://platform.ankra.app/api/v1/clusters/resources/roll-to \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"cluster_id": "<string>",
"version_id": "<string>"
}
'{
"ok": true,
"git_push": {
"status": "pushed",
"message": "<string>",
"repository": "<string>",
"branch": "<string>",
"commit_sha": "<string>",
"commit_url": "<string>",
"error_code": "GIT_PUSH_DEFERRED",
"kept_files": [
"<string>"
]
},
"commit_sha": "<string>",
"commit_url": "<string>",
"skipped_resources": [
{
"kind": "<string>",
"resource_id": "<string>"
}
],
"uninstalled_stacks": [
{
"resource_id": "<string>",
"name": "<string>",
"held_members": [
{
"resource_id": "<string>",
"kind": "<string>",
"name": "<string>",
"state": "<string>"
}
]
}
],
"resources_pending_delete": [
{
"resource_id": "<string>",
"kind": "<string>",
"name": "<string>",
"state": "<string>"
}
]
}{
"detail": "Cluster not found"
}{
"detail": [
{
"loc": [
"<string>"
],
"msg": "<string>",
"type": "<string>"
}
]
}Roll a cluster back to a resource version
Rolls a cluster’s resource definitions back to an earlier saved version, identified by cluster_id and version_id in the body, and moves the cluster’s current version to it. Requires clusters.operate; answers {"ok": true} on success, with skipped_resources, uninstalled_stacks and resources_pending_delete naming what the roll left as it was when there is any. On a cluster connected to a Git repository the rolled cluster is pushed to it once the roll is saved, to the repository and branch the cluster is connected to then (a roll never replays the cluster’s Git connection), and git_push says what the push did; commit_sha and commit_url name the commit when one landed. The roll is saved before it pushes, so a push that is deferred, fails or has nowhere to go still answers 200, with git_push.status, git_push.error_code and git_push.message saying what happened. The answer waits for the push: Ankra first tries to make sure the repository’s webhook exists and then pushes, each for at most two minutes, and then releases or keeps the hold the roll put on its add-on applies, for at most 30 seconds, so on a large cluster, behind another push holding the cluster’s push lock, or with a slow provider or database the answer can come up to about four and a half minutes after the roll is saved; a client that stops waiting does not stop the push, which still runs to its end.
import requests
url = "https://platform.ankra.app/api/v1/clusters/resources/roll-to"
payload = {
"cluster_id": "<string>",
"version_id": "<string>"
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({cluster_id: '<string>', version_id: '<string>'})
};
fetch('https://platform.ankra.app/api/v1/clusters/resources/roll-to', options)
.then(res => res.json())
.then(res => console.log(res))
.catch(err => console.error(err));curl --request POST \
--url https://platform.ankra.app/api/v1/clusters/resources/roll-to \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"cluster_id": "<string>",
"version_id": "<string>"
}
'{
"ok": true,
"git_push": {
"status": "pushed",
"message": "<string>",
"repository": "<string>",
"branch": "<string>",
"commit_sha": "<string>",
"commit_url": "<string>",
"error_code": "GIT_PUSH_DEFERRED",
"kept_files": [
"<string>"
]
},
"commit_sha": "<string>",
"commit_url": "<string>",
"skipped_resources": [
{
"kind": "<string>",
"resource_id": "<string>"
}
],
"uninstalled_stacks": [
{
"resource_id": "<string>",
"name": "<string>",
"held_members": [
{
"resource_id": "<string>",
"kind": "<string>",
"name": "<string>",
"state": "<string>"
}
]
}
],
"resources_pending_delete": [
{
"resource_id": "<string>",
"kind": "<string>",
"name": "<string>",
"state": "<string>"
}
]
}{
"detail": "Cluster not found"
}{
"detail": [
{
"loc": [
"<string>"
],
"msg": "<string>",
"type": "<string>"
}
]
}Authorizations
Response
Successful Response
Present when the cluster has a GitHub or Bitbucket Cloud repository row: what the push that followed the saved roll did. Absent for a cluster with no such repository, whose roll pushes nothing. The roll is saved whatever this says.
Show child attributes
Show child attributes
Present only when the target version held servers, Kubernetes nodes, bastions or VMs: a version roll never replays those, and each listed one keeps its current definition. Change them through their node group, control plane or VM settings.
Show child attributes
Show child attributes
Present only when the roll met uninstalled stacks. A version roll leaves a stack installed or uninstalled as it is, so each listed stack stays uninstalled. The roll wrote the version's definition to each of its held_members and deployed none of them; each is in the state shown. Down means off the cluster and stopping means still being removed from it; up, creating or updating means it may still be on the cluster, and the version changes nothing it deploys there. Installing the stack is what deploys its members again; a roll cannot.
Show child attributes
Show child attributes
Present only when the version has resources whose delete-for-good is still pending (state stopping with delete_permanently): the roll does not bring those back. They can be of any kind the roll replays, not only stack members: a stack description or a SOPS key, for example. Each listed resource took the version's definition and stays stopping, and the pending delete removes it. Once that delete has finished, rolling the cluster to the version again restores it.
Show child attributes
Show child attributes