Drumbeats is a monitoring service designed to solve the problem of silent failures in background jobs and services. Its core idea is heartbeat monitoring: each successfully running job actively sends a ping signal, and if no signal is received within the expected time, an alert is triggered. Unlike external polling and log scraping, heartbeat monitoring can detect issues that are hard to notice with polling and logs, such as jobs not running, running but failing silently, or running for too long.
Main capabilities:
- Supports multiple monitoring types: cron heartbeats, event-driven, and HTTP uptime monitoring, all managed in one place.
- Catches missed runs: if a job does not send a success ping within the schedule plus grace period, a MISSED alert is triggered.
- Catches delayed runs: running beyond the expected time triggers a grace period alert.
- Catches failed runs: shows exit codes, and failed runs surface.
- Catches hanging runs: alerts when a start ping is received but there is no corresponding success ping.
- Public status pages: turn a single monitor into a free public page with one click, or build a branded status page with components, incidents, and a maintenance timeline (Pro and above support custom branding, RSS, and subscribers).
- Notification channels: supports email, Slack, Telegram, Discord, Webhook, browser push, Microsoft Teams, and PagerDuty.
- AI-assisted setup: via the MCP protocol, you can integrate with AI coding assistants such as Claude Desktop, Cursor, VS Code, and Claude Code using natural language commands (such as "list my monitors"), with OAuth login and no API key required.
How to use:
- Simple integration: just one line of code per job, supports any language. Send an HTTP request to api.drumbeats.io/mcp or use HTTP requests directly.
- No SDK, agent, or library required.
- Can monitor jobs behind firewalls or in private VPCs because jobs send outbound HTTP requests and no inbound probes are needed.
- Works with any cron daemon, queue, or scheduler; the curl line stays the same when migrating from EC2 to Kubernetes.
- Free plan: up to 50 monitors, 200K Beats per month, minimum interval of 1 minute, ping history of 250 per monitor, maximum payload size of 100 KB, unlimited team seats, supports all notification channels.
- Pro plan: unlimited monitors, 1M Beats per month, minimum interval of 30 seconds, ping history of 1,000 entries/monitor, maximum payload size of 1 MB, custom domain for simple status pages, 1 branded page for advanced status pages (up to 25 monitors), up to 100 status page subscribers/page.
- Business plan: 3M Beats per month, ping history of 5,000 entries/monitor, maximum payload size of 2 MB, 3 branded pages for advanced status pages (unlimited monitors), up to 1,000 status page subscribers/page.
- All plans include unlimited team seats.
Use cases:
- Nightly database backups: if a backup job fails silently, a MISSED alert is triggered.
- SSL certificate renewal: a weekly heartbeat can detect a renewal that did not trigger before the certificate expires.
- Queue workers (such as Sidekiq, Celery, BullMQ): when a worker is killed due to blocked messages or OOM and produces no log output, a heartbeat can detect it within minutes.
- Scheduled crawlers and data syncs: send a success ping only when data is actually written, avoiding a false healthy state where an empty payload returns exit code 0.
- Payment and invoice reconciliation: when financial tasks such as Stripe payments, manual invoice batches, and dunning emails fail, notify the on-call person before the morning meeting.
- Other periodic tasks: any background job that needs to be ensured to run successfully on time.
Limitations:
- The free plan has a limited number of monitors (50), a limited Beats quota (200K/month), a minimum interval of 1 minute, and limited ping history.
- Advanced status pages, custom branding, RSS, and subscriber features require a Pro or Business plan.
- Sending pings needs to be integrated into the task code.
