Retry a site's failed deployment
Runs a deployment the settings mark can_retry again with the files it already built, and answers the same deployment with status queued. That is a failed deployment, and one other case: a deployment that was stopped (cancelled) in its after-deploy commands once its release was live, whose retry runs those commands again. Any other cancelled deployment is refused not_failed. A deployment that is not the site's newest would put an older release live, so it needs confirm_older true. Your role needs the Retry Deployments permission.
- Needs
deployas well: Always.
Path parameters
The id of a site row. A website's production site and each of its environments has its own id. Digits only. At least 1.
The id of a deployment of this site, as the settings list it. Digits only. At least 1.
Body
True to retry a deployment that is not the site's newest.
Returns
201 The deployment, queued again.
Fields of the 201 answer
One deployment of a site.
Fields
The deployment's current attempt. A retry moves it on.
The pipeline the deployment ran with.
True for a failed deployment, and for one stopped in its after-deploy commands once its release was live, unless it is a rollback, a stop of it is still putting the earlier release back, or its release archive was removed.
True when the site can roll back to this deployment's release.
True for a running deployment, and for a failed one whose scheduled tasks and queues stay paused after its database stage failed (crons_paused.state database_failed) while no other deployment of the site is in progress.
On a failed or stopped deployment whose crons and queues stay paused: state (database_failed or revert_unfinished) and message. A stop or a retry resumes them.
Why a failed deployment failed, or what a queued one waits for.
The commit's author (name, avatar_url), when the push said.
The commit's page at the provider, when the push said.
The commit's merge request or pull request (ref, kind, url), when the push said.
The path of the deployment's page in the Control Panel.
True for the deployment whose release is live.
On a failed deployment that switched maintenance mode on, whether it was switched off again (state, message).
The branch and commit a deployment deploys.
Fields
The first line of the commit's message.
The id of the release the deployment made. Null until its release goes live.
Who started it. Null for a push.
What a retry would do when this is not the site's newest deployment. Such a retry needs confirm_older.
The stage a running deployment is at, or a failed one stopped at.
waiting is a push that arrived while deploy on push is off; skipped is a waiting deployment that a later one overtook.
waitingqueuedrunningsucceededfailedcancelledskippedOn a stopped deployment whose running step could not be confirmed ended (state, checked_at).
The step a running deployment is at, or a failed one stopped at.
What started it. manual is Deploy now, webhook a push, rollback a rollback.
manualwebhookapirollbackunknownIf it is refused
| Status | Reason | What to do |
|---|---|---|
401 | none | Send a valid token as Authorization: Bearer <token>. |
403 | company_unavailable | Your company is unavailable. |
403 | forbidden | The person is not allowed to do this. |
403 | scope_required | Use a token with the scope named in scope. |
403 | two_factor_required | Turn on two-factor on your account, then send again. |
404 | feature_off | This feature is not on for your organisation yet. |
409 | artifact_pruned | The deployment's archive was removed after 30 days, so it cannot be retried or rolled back. Deploy again from the branch or commit. |
409 | deployment_running | A deployment of the environment is running, or a stopped one is still being cleaned up. Wait for it to finish, then try again. |
409 | not_failed | Only a failed deployment can be retried. Deploy again instead. |
409 | pipeline_changed_since_build | The pipeline changed after this deployment was built. Deploy again instead of retrying. |
409 | rollback_not_retryable | A rollback cannot be retried. Roll back again, or deploy. |
422 | confirm | The action needs an explicit confirmation field. Ask the person, then send it. |
429 | rate_limited | Wait for the seconds in Retry-After. |
Reasons are listed by name in Errors and refusals.
The same task elsewhere
curl -X POST https://manage.corefinity.com/api/corestack/sites/1042/deployments/5127/retry \
-H "Authorization: Bearer $COREFINITY_TOKEN" \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-d '{}'use Illuminate\Support\Facades\Http;
$response = Http::withToken(env('COREFINITY_TOKEN'))
->acceptJson()
->post('https://manage.corefinity.com/api/corestack/sites/1042/deployments/5127/retry', (object) []);
$data = $response->json();const res = await fetch('https://manage.corefinity.com/api/corestack/sites/1042/deployments/5127/retry', {
method: 'POST',
headers: {
Authorization: `Bearer ${process.env.COREFINITY_TOKEN}`,
Accept: 'application/json',
'Content-Type': 'application/json',
},
body: JSON.stringify({}),
});
// An answer with no body (204) has nothing to parse.
const text = await res.text();
const data = text && JSON.parse(text);import os
import requests
res = requests.post(
"https://manage.corefinity.com/api/corestack/sites/1042/deployments/5127/retry",
headers={
"Authorization": f"Bearer {os.environ['COREFINITY_TOKEN']}",
"Accept": "application/json",
},
json={},
)
data = res.json() if res.content else None{
"deployment": {
"attempt_id": 8035,
"binding": {
"id": 2211,
"label": "Default pipeline"
},
"can_retry": false,
"can_rollback": false,
"can_stop": false,
"crons_paused": null,
"duration_seconds": null,
"error": null,
"finished_at": null,
"git_author": null,
"git_commit_url": null,
"git_merge_request": null,
"href": "/resources/deployments/5127",
"id": 5127,
"is_current": false,
"maintenance_lift": null,
"queued_at": "2026-11-09T09:16:40Z",
"ref": {
"branch": "main",
"message": "Fix the basket count in the mobile header",
"sha": "9730595c6ec7aa59b77674ada74e30d961b1817c",
"short_sha": "9730595"
},
"release_id": null,
"requested_by": {
"id": 3301,
"name": "Daniel Reyes"
},
"retry_warning": null,
"stage": null,
"stage_label": null,
"started_at": null,
"status": "queued",
"status_label": "Queued",
"stop_termination": null,
"task": null,
"trigger": "manual"
}
}{
"message": "Unauthenticated."
}{
"code": "company_unavailable",
"message": "Your company is unavailable.",
"reason": "company_unavailable"
}{
"code": "feature_off",
"message": "Not found.",
"reason": "feature_off"
}{
"code": "artifact_pruned",
"message": "This deployment's archive was removed after 30 days, so it can't be retried or rolled back. Deploy again from the branch or commit.",
"reason": "artifact_pruned"
}{
"code": "confirm",
"errors": {
"confirm_older": [
"This will put an older release live: commit 6b34ba4 from 2 Nov at 14:05. The newest deployment is 49ea0cf from 9 Nov at 09:16."
]
},
"message": "This will put an older release live: commit 6b34ba4 from 2 Nov at 14:05. The newest deployment is 49ea0cf from 9 Nov at 09:16. To retry it anyway, repeat the request with confirm_older set to true.",
"reason": "confirm"
}{
"code": "rate_limited",
"message": "Too many requests. Try again in 60 seconds.",
"reason": "rate_limited",
"retry_after_seconds": 60
}Try it is off for this endpointIt runs commands on a live site. Copy the request and run it yourself when you are sure.