A permission changes, a token expires, the AI credits run out. The automation stops and nobody is told. Automation Monitor gives each critical automation a dead-man's switch and alerts the right person the moment an expected run doesn't happen.
Three ways it usually goes. The failure is quiet, the discovery is loud, and the gap between them is where the damage happens.
Every night an automation pushes approved invoices to the accounting integration. The integration's token expires. Nothing errors in a way anyone sees. Invoices quietly stop going out until accounts payable asks why nothing has arrived.
An agency builds lead routing for a client. Six weeks on, the client deactivates the user who set it up, and every automation that ran as that user stops with them. The sales team sees an empty pipeline and blames the agency.
AI-assisted triage tags and assigns incoming requests. On the 19th the account's AI credits run out. Over the weekend, requests pile up untagged and unassigned, and the Monday morning queue is a wall.
Automation Monitor doesn't need a run-history API or access to your boards. The automation itself reports in, and we notice when it goes quiet.
Open any automation you care about and append one action from the Monitor block. Save it. The next time the automation runs, the heartbeat shows up in your Fleet inbox, ready to name.
Every N hours, or on a business-hours schedule in the team's time zone. Add a grace period so a slow run doesn't page anyone. That expectation is the monitor.
A missed heartbeat opens an incident and alerts the people you chose, in monday, by email, or in Slack. The next successful run resolves it and sends a recovery alert, so nobody has to go and check.
If the problem is that monday can't reach us, or we can't reach monday, a fallback email tells you that too.
One action appended to any automation. If the expected heartbeat doesn't arrive, that's the alert.
Every N hours, or 09:00 Mon–Fri in the team's time zone. Weekends and holidays don't raise false alarms.
A slow run isn't a failure. Set how long to wait before an incident opens.
The next successful run resolves the incident and tells the same people it's over.
Scheduled checks of actual board state, like "no item in Pending for more than 4 hours". They catch failures in automations nobody instrumented.
Route each monitor to the people who can fix it. Different teams, different channels.
We monitor our own connection to your account. If monday alerting itself breaks, you still hear about it.
A board view with the health of every monitored automation, open incidents first.
A heartbeat from an automation you haven't named yet lands in the inbox with a guided setup. No forms to fill in first.
When it failed, when it recovered, who acknowledged it. 7 days, 90 days or 12 months depending on plan.
Acknowledge an incident so the team knows it's being handled. Pause a monitor during planned changes.
Admins and members manage monitors. Viewers and guests see the Fleet view read-only.
| State | Monitor | Last signal | Expectation | Actions |
|---|---|---|---|---|
| Alerting | Invoice syncFinance ops · heartbeat | Expected 09:00 · 1 h 12 min late | every 24 h · 15 min grace | |
| Late | Daily standup digestEngineering · heartbeat | Expected 09:00 · in grace, 6 min | 09:00 Mon–Fri · Europe/London | |
| New | Heartbeat from "Client onboarding"Received 2 min ago · needs a name and cadence | First heartbeat 2 min ago | not set | |
| Healthy | Lead routingSales pipeline · heartbeat | Last run 4 min ago | every 1 h · 10 min grace | |
| Passing | Nothing stuck in Pending > 4 hSupport queue · assertion | Checked 11 min ago · 0 items | every 30 min | |
| Healthy | Onboarding checklistPeople ops · heartbeat | Last run 22 min ago | every 4 h · 20 min grace | |
| Paused | Refund approvalsFinance ops · heartbeat | Paused by Priya · board migration | every 2 h · 10 min grace |
Nothing here. All monitors are healthy.
You own the account, so every silent failure lands on your desk eventually. Put a heartbeat on the automations that move money, customers or deadlines, and hear about failures first.
Your processes run on automations other people built. A business-hours schedule with a grace period gives you a monitor that matches how the work actually flows, with no false alarms at 3 a.m.
You built it, so you're the one who gets the call. Route each client's monitors to your own Slack channel and be the one who tells them, with the board name and the last successful run already in hand.
Start free. Upgrade inside monday.com when you need more monitors, longer history or more places to send alerts.
Billing happens inside monday.com on your existing account. Paid plans start with a 14-day trial. Prices in USD.
Anything else: support@automationmonitor.app. We answer within two business days.
No. Heartbeats don't read your boards at all. The automation sends us a signal, and we only store the monitor's name, the board it came from and when it ran.
Assertions are the exception: to check that no item is stuck in Pending, we have to read that board. We only read the boards you add assertions to.
Your access tokens are revoked immediately. Everything stored for your account, including monitors, incidents, alert settings and heartbeat history, is deleted within 7 days.
No. It's one extra action at the end of the automation, and it doesn't block the actions before it. Like any other step, it counts as one automation action in monday.
One heartbeat expectation or one assertion. An automation with a heartbeat step is one monitor however often it runs.
We monitor our own connection to your monday account. If our access stops working, so would monday notifications, so we email the fallback address you set and tell you how to reconnect.
Through monday.com, on the account you already have. Paid plans start with a 14-day trial, and you can change or cancel the plan from the monday marketplace at any time. We never see your card details.
Anyone who can open the board you add it to. Admins and members can set up monitors, acknowledge and pause. Viewers and guests get a read-only view, which is handy for client stakeholders.
Access uses monday's OAuth, scoped to what each feature needs. Tokens are encrypted at rest, all traffic is TLS 1.2 or better, and data is deleted within 7 days of uninstall. The privacy policy has the detail.
Three monitors are free, forever. Add a heartbeat to the automation you'd least like to discover had stopped.