For developers
Uptime Monitoring for Developers
A full API and CLI from the free plan, JSON assertions on responses, heartbeats for the jobs with no URL, and an MCP server so your assistant can run it. Monitoring that doesn't make you click.
Why developers pick Sentinel
Monitoring you can script
Most uptime tools hide the useful parts behind the dashboard, then gate the API on the expensive plan. Sentinel does the opposite. Everything the dashboard can do, the REST API, the CLI and the MCP server can do too, from the first free account.
- Create, update and pause monitors from code on any plan
- Assert on the JSON an endpoint returns, not just its status code
- Heartbeats for cron jobs, queues and anything without a URL
Made for people who'd rather script it
The dashboard is optional
A real API, from day one
Full read and write access on every plan, Free included. Tokens carry scoped permissions, responses are plain JSON, and the reference documents every endpoint with examples.
The CLI
Install @rootstuff/sentinel from npm and add a monitor, pause it, or pull a report from the terminal. Handy in a deploy script, a Makefile, or a quick check before you leave for the day.
Works with your AI assistant
Sentinel ships an MCP server. Connect it to Claude, Cursor or any MCP client and ask what's down, create a monitor for the service you just shipped, or pull a week of response times, in plain language.
JSON assertions
A 200 from a health endpoint that says status: degraded is still a failure. Assert on fields in the response body with equals, contains, greater-than and regex, and alert when any assertion fails.
Heartbeats and cron
Give a cron job, a queue worker or a backup script a URL to ping when it finishes. If the ping doesn't arrive on schedule, Sentinel alerts you. Grace periods and cron expressions are supported.
Signed webhooks
Every alert can post signed JSON to an endpoint you own. Deliveries are logged with the response you returned, so when your handler misbehaves you can see exactly what it got.
Is the API available on the free plan?
Yes, with full read and write access. A free account can create, update, pause and delete monitors through the REST API, the CLI, Terraform or the MCP server. Writes on free accounts are rate limited per hour, which is plenty for scripts and well short of abuse.
What is a heartbeat monitor?
A monitor that expects to be pinged rather than one that does the pinging. You give a scheduled job a unique URL and have it request that URL when it completes. If the request doesn't arrive within the expected interval plus a grace period, Sentinel opens an incident and alerts you. It's the simplest way to know a cron job silently stopped.
Can I check the body of an API response, not just the status code?
Yes. Add JSON assertions to an HTTP monitor using dot-notation paths into the response, with operators like equals, not_equals, contains, exists, greater-than and regex. Up to ten assertions per monitor, and the monitor fails when any of them does.
How do I use Sentinel from Claude or another AI assistant?
Connect the MCP server. It's listed in the Claude Connectors Directory, and any client that speaks MCP over HTTP can use it with an API token. The server exposes 26 tools covering monitors, incidents, groups, uptime summaries, performance history and on-demand URL checks.
Is there a Terraform provider?
Yes, on the Terraform Registry, covering monitors, groups and notification settings. It's the right fit once monitoring belongs in the same repository as the infrastructure, and it's included on the Business plan.
What languages or frameworks does it support?
Any. Sentinel checks over HTTP, TCP, ICMP and DNS from the outside, so it doesn't care what the service is written in and there's nothing to install in your app. The only thing it asks of your code is an optional ping to a heartbeat URL.
Monitoring that fits in a shell script
Grab a token, run one command, and your first monitor is live. Free plan, full API, no card.
No card for the free plan · 14-day trial on paid plans · Cancel anytime