Quiet failures cost real money.
When a cart bug flatlines your orders or a nightly report stops sending, nothing throws an error. SignalSitter watches the signals your business runs on, so you know something's wrong long before your customers do.
✓ 10 rule types, statistical included
✓ Backtest on past data
✓ Agent-native over MCP
✓ Setup-free monitors
Silent failures
The failures that hurt most don't throw errors.
Your scheduled job never fires. A bug in your shopping cart flatlines your sales. A server stops reporting. Nothing crashes, so nothing tells you. SignalSitter does.
nightly-backup · nothing since 02:00
It went quiet.
Tell your jobs and services to send SignalSitter a simple HTTP request when they run. If one is late, past the grace period you set, we let you know immediately.
payment_errors · 14× its recent average
It spiked.
Send the numbers that matter with each heartbeat. When errors or refunds shoot past their usual level, an incident opens on the spot.
signups · well under recent average
It drifted.
Some problems creep in. Statistical rules compare each reading with its own recent average, trend and spread. You set how far is too far.
Rules
Stop getting woken up for nothing.
Alert rules are usually fiddly to write and noisy once they're on. Here, you or your agent write one in seconds, then backtest it on your recent history before it can wake anyone.
Would have fired twice: day 9 and day 23. Every incident shows the reading and the rule it broke.
Agent-native
No time to set up monitoring? Hand it to your agent.
SignalSitter is agent-native to its bones. Everything you can do in the dashboard or the REST API, your agent can do too over remote MCP, faster and without you reading the docs.
Setup-free monitors
Send your first heartbeat or measurement, and SignalSitter creates the monitor on the fly.
Incidents handled, not just reported
Your agent can look into an incident, acknowledge it and resolve it the moment it opens.
Docs your agent reads for you
The whole manual is one fetch away, in a format agents read well.
Alerts
An alert nobody sees is one more silent failure.
Send each monitor's alerts where your team already looks.
Any webhook
Your own code, Zapier, Make, n8n or your on-call tool. Every alert is signed, so your code can tell it came from us.
customer-report-batch: report_count is 0. It must never be zero.
Your team, where they already are
Slack, email or Telegram, to one teammate or everyone on a project.
Pricing
Every rule on every plan. Pay for how much you watch.
Prices in US dollars per month, excluding applicable taxes. One plan covers your organization, and every project and teammate shares it.
Free
$0
No card needed
1 active monitor
As often as every 5 minutes
2 measurements per beat
Readings kept for 14 days
1 callback endpoint
Get started
Builder
$39
per month
25 active monitors
As often as every minute
8 measurements per beat
Readings kept for 90 days
5 callback endpoints
Get started
Smart
$199
per month
100 active monitors
As often as every minute
16 measurements per beat
Readings kept for 365 days
20 callback endpoints
Get started
Scale
$999
per month
500 active monitors
As often as every minute
16 measurements per beat
Readings kept for 365 days
50 callback endpoints
Get started
Additional monitors cost $2 per monitor per month, up to 2,000 monitors in total, including above Scale.
Your allowance is your plan’s included monitors plus the additional monitors you buy.
API and MCP request rates scale with your total allowance. Callback endpoint slots scale up to 50.
History, interval and measurements per beat stay at your plan’s limits.
Readings are kept for the number of days shown on your plan.
What happens at my monitor limit?
You can’t create another until you buy more capacity, delete or archive one, or upgrade. Monitors you already have keep running.
What if I send too many requests?
Requests over your plan’s rate get a rate-limit response that says when to retry. Keep going over after a warning and ingest pauses until you resume it. Nothing is deleted.
Are there overage charges?
No automatic overage charges. You choose and pay for additional monitor capacity before using it.
What if I move to a smaller plan?
Your oldest monitors keep running and the rest pause until there’s room. Their history stays, and nothing is deleted.
FAQ
Can it do that? Yes.
Can it catch a job that ran on time but did nothing?
Yes! Send the count it produced with each heartbeat. A "never zero" or minimum rule flags the empty run.
Can it watch business numbers, not just jobs?
Yes! Orders per hour, signups, refunds against sales: anything you can count can carry a rule.
Can I test a rule before it goes live?
Yes! Backtest it on your recent history and see every time it would have fired.
Can my agent run the whole thing?
Yes! Everything in the dashboard and the API is available to your agent over MCP, from setup to closing incidents.
Can the first heartbeat create its own monitor?
Yes! With a manage key, headers on that first heartbeat set the monitor's name, how often it should arrive and how late it may be.
Can I send alerts to Zapier, Make or n8n?
Yes! Point the signed webhook at their webhook trigger and build whatever happens next.
Can I verify that an alert really came from you?
Yes! Every webhook is signed with a secret only you and SignalSitter hold, so your code can check it.
Can I keep my data in the EU?
Yes! Choose EU residency for any monitor, and its data stays in the EU.
Ready when you are
Send your first heartbeat in seconds.
Privacy: EU data residency for any monitor you choose. SignalSitter does not use an analytics provider, session recording, or a model provider for customer monitoring data.
Privacy information →