Health check

Alertee runs your checks on a cron engine. If that engine stalls, no checks run and no alerts fire — so it is the critical path. The /health endpoint exposes the engine’s liveness so you can watch it from an external uptime monitor (UptimeRobot, Better Stack, Pingdom, etc.).

Endpoint

GET https://prod-neon-alertee-gru2.encr.app/health

No authentication. It returns only timestamps and counts — no customer data.

200 OK — the every-minute cron completed a run within max_age_seconds.
503 Service Unavailable — the cron has not completed a run in time (engine stalled, slow, or never started). The JSON body is still returned.

The status code is what your uptime monitor should alert on.

Tuning the threshold

The default staleness threshold is 180 seconds (the engine ticks every 60s, so this tolerates a couple of missed/slow ticks). Override it per monitor:

GET /health?max_age_seconds=120

Response

{
  "status": "ok",
  "now": "2026-06-02T12:00:05Z",
  "max_age_seconds": 180,
  "crons": [
    {
      "job": "run-checks-team",
      "last_run_at": "2026-06-02T12:00:00Z",
      "seconds_ago": 5,
      "last_duration_ms": 1820,
      "last_due_count": 12,
      "stale": false
    }
  ]
}

status is driven by the run-checks-team cron (the every-minute job — the tightest heartbeat). The slower run-checks-starter and run-checks-free jobs appear in crons for diagnostics but do not change the overall status. When status is down, a message field explains why.

Wiring up UptimeRobot

Add monitor → HTTP(s).
URL: https://prod-neon-alertee-gru2.encr.app/health
Monitoring interval: 1 minute.
Alert on a non-2xx status code (the 503 is your signal).

Getting started

Checks

Inbox

Alerts

Account

Status

Endpoint

Tuning the threshold

Response

Wiring up UptimeRobot

​Endpoint

​Tuning the threshold

​Response

​Wiring up UptimeRobot

Endpoint

Tuning the threshold

Response

Wiring up UptimeRobot