GET /statusReturns a JSON object describing the current state of all monitored services
in the requested region (or all regions if region is omitted).
The HTTP status code mirrors the health field: 200 when healthy,
503 when unhealthy. This allows simple curl -f or monitoring-tool
checks without parsing JSON.
| Parameter | Type | Default | Description |
|---|---|---|---|
region | string | (aggregated) | Scope the response to a single region, e.g. nsc-nbg. When omitted, services from all configured regions are included. |
ignore_downtimes | boolean | false | Suppress any active scheduled downtime — downtime is null and health reflects the true service state. Useful for automation that must not be fooled by patch-day windows. |
{
"health": true,
"downtime": null,
"outage": null,
"last_checked": "2026-09-25T07:36:39+00:00",
"last_good": "2026-09-25T07:36:39+00:00",
"services": [
{
"key": "compute",
"label": "Compute",
"ok": true,
"timeout": 60,
"last_checked": "2026-09-25T07:36:39+00:00",
"last_good": "2026-09-25T07:36:39+00:00"
},
// ...
]
}
| Field | Type | Description |
|---|---|---|
health | boolean | true when at least one service is passing and none are failing (or an active downtime suppresses failures). Determines the HTTP status code: 200 vs 503. |
downtime | object | null | The currently active scheduled downtime, or null. See below. |
outage | object | null | The current ongoing outage, or null. See below. |
last_checked | string | null | ISO-8601 UTC timestamp of the most recent poll across all services. |
last_good | string | null | ISO-8601 UTC timestamp of the last time any service was healthy. |
services | array | Per-service status objects (see below). |
| Field | Type | Description |
|---|---|---|
key | string | Machine-readable service identifier (e.g. compute, object_store). |
label | string | Human-readable name. |
ok | boolean | null | true=passing, false=failing, null=unknown/not applicable. |
timeout | integer | Seconds after which a service is considered failing. |
last_checked | string | null | ISO-8601 UTC timestamp of the last poll. |
last_good | string | null | ISO-8601 UTC timestamp of the last healthy poll. |
{
"start": "2026-09-25T10:00:00+00:00",
"end": "2026-09-25T12:00:00+00:00",
"text": "Patchday: upgrading compute nodes",
"created_at": "2026-09-24T15:30:00+00:00"
}
During a downtime, the dashboard fades indicators and shows "Maintenance".
Automation should use ignore_downtimes=true to see the real health state.
{
"service": "object_store",
"started_at": "2026-09-25T07:30:00+00:00",
"reason": "Object Store: [write] ConnectionRefused...",
"ended_at": null,
"duration_sec": null,
"ongoing": true
}
An outage is created automatically when a service transitions from healthy to
failing. It is closed (filled with ended_at and duration_sec) when the
service recovers.
Since /status returns HTTP 503 when unhealthy, curl -f suffices — no
JSON parsing needed:
curl -sf "https://status.noris.cloud/status?region=nsc-nbg&ignore_downtimes=true" \
|| (echo "ABORT: OpenStack status not good" && exit 1)
The -s suppresses progress, -f fails on HTTP errors (including 503).
The || branch prints a message and returns a non-zero exit code.
To inspect the response body on failure:
curl -sf "https://status.noris.cloud/status?region=nsc-nbg&ignore_downtimes=true" \
| jq -e '.health == true' \
|| (echo "ABORT: OpenStack status not good" && exit 1)
Because the endpoint returns 503 on unhealthy, the uri module fails
automatically — no conditional needed:
- name: Check OpenStack status before proceeding
uri:
url: "https://status.noris.cloud/status?region=nsc-nbg&ignore_downtimes=true"
status_code: 200
If you need the response body for diagnostics:
- name: Check OpenStack status before proceeding
uri:
url: "https://status.noris.cloud/status?region=nsc-nbg&ignore_downtimes=true"
return_content: true
status_code: 200
register: os_status
- name: Abort if OpenStack is unhealthy
fail:
msg: "OpenStack status is not healthy — aborting play"
when: not (os_status.json.health | default(false))
The ignore_downtimes=true parameter ensures the check sees real health even
during a scheduled patch-day window.
During planned maintenance (patch days), a downtime window is set with a
start, end, and descriptive text. While active:
/status downtime field contains the schedule and description.health field reports true — failures during an active downtime
are expected, so health is suppressed to OK.ignore_downtimes=true to see the real service state during a
downtime window (automation that must not be fooled by patch days).| Method | Path | Auth | Description |
|---|---|---|---|
GET | / | none | HTML dashboard (with region tabs) |
GET | /regions | none | List available region names (["nsc-nbg"]) |
GET | /status | none | JSON status (↓ region, ignore_downtimes) |
GET | /healthz | none | Liveness probe ({"ok": true}) |
GET | /outages | none | Recent outage history (↓ region) |
GET | /version | none | Git commit hash (plain text) |
GET | /downtimes | none | List scheduled downtimes (↓ region) |
↓ = accepts the query parameter