Drumbeats

About Drumbeats

Drumbeats is heartbeat monitoring for the failures external probes cannot see. Background jobs break in three ways — they never run, they run and fail silently inside a swallowed exception, or they hang forever on a stuck connection — and none of those show up in an HTTP check against your homepage or in a log aggregator that only knows what your code thought to report. Drumbeats inverts the model: every successful run announces itself with a single curl, each monitor expects that ping inside a known window, and the absence of the announcement is the alert. Missed runs are explicit rather than inferred, late runs trigger grace-period alerts, failed runs surface the exit code, and a hung run is caught when the start ping arrives without its matching success ping. Because the job dials outward, it works behind firewalls with no inbound probing, and it survives infrastructure changes — moving from EC2 to Kubernetes leaves the curl line untouched — and it is indifferent to which cron daemon, queue or scheduler you use. Cron, heartbeat, queue-worker and HTTP uptime monitors all live in one place with incidents that open themselves, including expiring TLS certificates, and eight alert channels: email, Slack, Telegram, Discord, webhook, browser push, Microsoft Teams and PagerDuty. Every monitor gets a public page, and a REST API plus an MCP server let you create, ingest and resolve programmatically or point an agent at it. Billing is on activity rather than monitor count, and seats are never charged.

Key Features

Tags