Long-running work
Changes that take time answer an operation, which you read until it finishes.
On this page
A change that cannot finish within the request answers an operation. Read the operation until its status is completed or failed.
What a change answers
| Status | Body | Meaning |
|---|---|---|
201 | An operation, or an object with an operation field | The work has started |
202 | An operation, or an object with an operation field | The work is accepted. The operation's status says whether it has started |
200 | "operation": null | Nothing differed, so nothing was started |
201 | "queued": true and "operation": null | The change waits for a window; see Apply now or in a window |
A Varnish change answers 202 both when it has started and when it waits for a window. Check queued first, then the operation's status.
Read an operation
GET / needs the read scope. Reading checks progress first when the last check is a few seconds old, so one request every few seconds is enough.
curl https://manage.corefinity.com/api/corestack/operations/77120 \
-H "Authorization: Bearer $COREFINITY_TOKEN" -H "Accept: application/json"{
"id": 77120,
"type": "php_version",
"status": "completed",
"from_version": "8.2",
"to_version": "8.3",
"steps": [{ "key": "saved", "label": "Saved", "state": "done" }],
"error": null,
"created_at": "2026-10-24T09:12:03Z",
"completed_at": "2026-10-24T09:14:41Z"
}The answer carries more fields than shown here; the operation's reference page lists them all. Applying pending changes now answers a shorter operation of its own shape.
Statuses
status | status_label | Final |
|---|---|---|
pending | Saving change | No |
committed | Waiting for the platform | No |
rolling_out | Applying change | No |
executing | Running | No |
completed | Done | Yes |
failed | Failed | Yes |
While an operation is not final, another change to the same site is refused 409 busy.
A new site or environment can wait before its work starts: its status stays pending and its status_label reads Waiting for the platform or Waiting for DNS. It carries on by itself, so do not create it again.
When an operation fails
error says why in plain words, and failure_reason carries a stable code where there is one.
Not every failed operation can be retried: read its can_retry. While it is true, POST / carries on from where the operation stopped.
When retry_requires is acknowledge_env, check the site's .env file first and send "acknowledge_env": true with the retry.
While can_retry is false, a retry is refused: make the change again. The retry's reference page lists its refusals and what to do about each.
A failed site or environment creation is never retried this way, and the retry answers 422 not_retryable: ask Corefinity to retry it, or delete the site or environment and create it again.
Results shown once
An operation never carries a secret. Where work creates a login, POST / answers it once, with the exec scope; a second request is refused 409 already_revealed.
Work followed on its own endpoint
Most changes answer an operation read with GET /. Some areas follow their work on an endpoint of their own, which the request's reference page names.
Read it the same way: every few seconds, until its status is final.
| Read this | What it does |
|---|---|
GET / | Get an operation |
GET /enterprise-copies/{copy} | Get an Enterprise copy |
GET / | Get a deployment |
GET /performance-audits/{audit} | Get a performance audit |
GET / | List an audit's GTmetrix reports |
GET / | Get an access operation |
GET /corestack/commands/{command} | Get a command |
GET / | Get a database operation |
GET / | Get a rehearsal |
GET / | Get an email operation |
GET / | Get an error pages operation |
GET /firewall/changes/{change} | Get a firewall change |
GET / | Get a logs export |
GET / | Get a CoreCluster's create progress |
GET / | Get an Nginx rules operation |
GET / | Get a site's PageSpeed history |
GET / | Get an environment's host keys |
GET /ssh/changes/{change} | Get an SSH change |
GET / | Get a Varnish operation |
A snapshot is followed on its own endpoint too: read GET /corestack/snapshots/{snapshot} until its status is ready or failed.