Monitor the platform's status JSON

Poll the public status JSON the website serves, read each feature's state and freshness, and alert on what matters, the way the studio's own uptime monitor does.

Developers and AI

The public website serves the platform's state as JSON at https://www.blockoutgames.com/api/status: the API, the database, the collector and its jobs, the assistant and push notifications, as booleans and times only. This recipe polls it from any uptime monitor and alerts on the same checks the studio's own monitor uses.

Before you start

  • An uptime monitor that can fetch a URL and check a JSON field: Gatus, Uptime Kuma, Better Stack or a script on a timer.

  • Nothing else. The status needs no sign-in or key, and names no host, address or error text.

What /api/status answers

GET/api/status

200 with the platform's state, or 503 with Retry-After: 30 and { "error": "The platform did not answer" } when the website cannot reach the API, so every check fails.

JSON
{  "checkedAt": "2026-10-27T09:14:30.000Z",  "api": { "ok": true },  "database": { "ok": true },  "collector": { "state": "ok", "lastSeenAt": "2026-10-27T09:14:21.000Z" },  "jobs": {    "snapshot": { "fresh": true, "lastOkAt": "2026-10-27T09:14:00.000Z" },    "sorts": { "fresh": true, "lastOkAt": "2026-10-27T09:10:02.000Z" },    "recommendations": { "fresh": true, "lastOkAt": "2026-10-27T09:12:40.000Z" },    "creators": { "fresh": true, "lastOkAt": "2026-10-27T09:05:11.000Z" },    "github": { "fresh": true, "lastOkAt": "2026-10-27T08:50:00.000Z" }  },  "assistant": { "connected": true },  "push": { "ok": true }}

Status

checkedAtstring
When the API last worked the status out, ISO 8601. It does so at most every 30 seconds.
api.okboolean
The API answered and reached its database; today always the same value as database.ok.
database.okboolean
The API reached the database.
collector.statestring
ok when the collector's heartbeat is under a minute old and reports its numbers, stale when it is older or reports none, down with no heartbeat at all.
collector.lastSeenAtstring | null
The collector's last heartbeat.
jobs.<job>.freshboolean
The job succeeded recently enough: snapshot within 15 minutes, recommendations 20, sorts and creators 30, github 60. About three of each job's own intervals.
jobs.<job>.lastOkAtstring | null
The job's last success within the collector's last hour of runs.
assistant.connectedboolean
The assistant's service answered its readiness check.
push.okboolean
Push is set up and delivered within the last 5 minutes (it runs every 15 seconds).

Steps

  1. Add the endpoint

    Point the monitor at https://www.blockoutgames.com/api/status, every 60 seconds. The answer changes at most every 30 seconds, so polling faster tells you nothing new.

  2. Check the status code

    Alert on anything but 200: the platform did not answer.

  3. Check the fields that matter

    The studio's own monitor checks collector.state is ok, jobs.snapshot.fresh, jobs.github.fresh, assistant.connected and push.ok. Add api.ok and database.ok.

  4. Give stale a grace period

    A stale collector or a job that is not fresh is late, not down. Alert after two or three failed checks in a row rather than the first.

  5. Point people at the page

    Link the alert to the status page, which shows the same features with their history.

yaml
endpoints:
  - name: blockout-platform
    url: https://www.blockoutgames.com/api/status
    interval: 60s
    conditions:
      - '[STATUS] == 200'
      - '[BODY].collector.state == ok'
      - '[BODY].jobs.snapshot.fresh == true'
      - '[BODY].jobs.github.fresh == true'
      - '[BODY].assistant.connected == true'
      - '[BODY].push.ok == true'

The live stats JSON

https://www.blockoutgames.com/api/stats answers what the platform has collected, for the website's live numbers: metric points stored (points.total, perSecond, lastHour, lastDay), games, genres, creators and sorts tracked, readings a minute and the last one, players now in the tracked games read in the last hour, GitHub repositories and pull requests, the oldest reading kept, and collector.state. It is a copy at most 10 seconds old, cacheable for 10 seconds, and 503 with Retry-After: 30 when unavailable. /api/stats/stream sends the same as Server-Sent Events every 10 seconds, up to 4 streams per address. Use /api/status for monitoring; the stats are counts, not checks.

What you get

An alert within a couple of minutes when the API, the database, the collector, one of its jobs, the assistant or push stops working, from a public endpoint that costs the platform nothing extra: the API works the status out at most every 30 seconds however many monitors ask.

Updated