Traditional APM daemons chew up hundreds of megabytes of RAM just to miss a silent exit code. PingWhip monitors periodic tasks, queue workers, and batch pipelines with a single HTTP ping. No agents, zero daemon bloat.
No dependencies to compile. No services to register. Pure HTTP standard.
#!/usr/bin/env bash
set -Eeuo pipefail
TOKEN="pw_live_8f3a92b1c4"
START_TIME=$(date +%s)
# Trap unhandled runtime errors and exit immediately to /fail
trap 'curl -fsS -m 5 -X POST "https://pingwhip.com/ping/${TOKEN}/fail" \
-H "Content-Type: application/json" \
-d "{\"error\": \"Exit $? on line $LINENO\", \"duration\": $(($(date +%s) - START_TIME))}"' ERR
# Run critical execution payload
python3 /srv/workers/reconcile_ledgers.py
# Report successful completion heartbeat with execution duration
DURATION=$(($(date +%s) - START_TIME))
curl -fsS -m 5 "https://pingwhip.com/ping/${TOKEN}?duration=${DURATION}"
{
"status": "acknowledged",
"edge_pop": "EWR (Newark, NJ)",
"telemetry": {
"duration_recorded": 14,
"ip_hash": "sha256:7e8b91a24..."
},
"monitor": {
"slug": "reconcile-ledgers",
"state": "HEALTHY",
"grace_period": "900s",
"next_expected_at": "2026-11-21T03:15:00Z"
}
}
Decoupled edge telemetry built on Cloudflare Workers and PostgreSQL. We completely separate ping ingestion from state evaluation so your cron tasks never block.
Terminates at the nearest Cloudflare Anycast edge across 300+ PoPs worldwide. Validates SHA-256 tokens in memory and returns a 200 ACK with zero database lock contention.
Distributed cron runner evaluates Postgres state every 60s (10s on Pro). When the current UTC timestamp exceeds interval plus grace buffers, the monitor state shifts to DOWN.
Parallel asynchronous queues dispatch failure events across transactional email (Resend) and authenticated Webhooks (Slack, Discord, Opsgenie, custom HTTP) scoped per monitor.
Predictable limits. No per-seat pricing. Zero surprise overages.