Recipes

Copy-paste integrations for the places alerts usually come from.

GitHub Actions: page on failed deploy

- name: Page on-call on failure
  if: failure()
  run: |
    curl -sS -X POST https://anypager.app/api/v1/alerts \
      -H "Authorization: Bearer ${{ secrets.ANYPAGER_TOKEN }}" \
      -H "Idempotency-Key: gha-${{ github.run_id }}" \
      -H "Content-Type: application/json" \
      -d "{\"message\": \"Deploy failed: ${{ github.repository }} @ ${{ github.ref_name }}\", \"severity\": \"critical\"}"

The Idempotency-Key makes reruns safe — a retried job will not page twice. Store the ap_… token as an encrypted repository secret.

Cron health check

#!/bin/sh
set -eu

if ! curl -fsS --max-time 10 https://example.com/health > /dev/null; then
  curl -sS -X POST https://anypager.app/api/v1/alerts \
    -H "Authorization: Bearer $ANYPAGER_TOKEN" \
    -H "Idempotency-Key: health-$(date -u +%Y%m%d%H%M)" \
    -H "Content-Type: application/json" \
    -d '{ "message": "example.com health check failed", "severity": "critical" }'
fi

The minute-granular idempotency key collapses repeated failures within the same minute into one page.

Any shell script, via the CLI

If the machine has the CLI paired, skip curl entirely:

./backup.sh || anypager --critical "nightly backup failed on $(hostname)"

Long-running job notifier

if make release; then
  anypager "release build done"
else
  anypager --critical "release build FAILED"
fi

The explicit if keeps the two outcomes separate — with make release && notify || page, a failed success-notification would also fire the failure page.

Webhook bridge

Anything that can send an HTTP request can page you — Grafana, Uptime Kuma, Alertmanager, a Stripe webhook handler. Point it at POST /api/v1/alerts with the token in the Authorization header, map your payload's text into message, and set severity to critical for the pages that must wake someone up.