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.
SIGNALS BEFORE MONITORS
FIXED-THRESHOLD RULES
AGENT-NATIVE OVER MCP
SIGNED, RETRIED WEBHOOKS
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.
Choose a missing-beat heartbeat period on your Signal, or disable it. If publishing stops beyond that explicit period, the Signal sends its missing-beat callback.
payment_errors · above a fixed maximum
It crossed the maximum.
Publish payment_errors to a Signal. A linked Monitor checks it on its schedule and opens an incident when the value crosses your fixed maximum.
signups · below a fixed minimum
It fell below the minimum.
Publish signups to a Signal. Set a fixed numeric minimum on a linked Monitor so a scheduled check catches readings below that limit.
Rules
Stop getting woken up for nothing.
Create a Signal first: it owns the publish URL and readings. Then link a Monitor to check that shared Signal on its schedule, using clear numeric limits. Only deterministic fixed-threshold rules are currently available.
A Monitor checks the shared Signal on its schedule. Values above 5 open an incident with the reading and fixed limit.
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.
Create a Signal first
A first publish with a manage key may create a Signal, never a Monitor. Then link one or more Monitors, each with its own rules, schedule, incidents and callback.
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 Signal missing-beat callbacks and Monitor incident callbacks 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. Its fixed minimum is 1.
Your team, where they already are
Slack, email or Telegram, to one teammate or everyone on a project.
Pricing
Fixed thresholds 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 Signal · 1 Monitor · 8,640 pooled stored events per 30 days
As often as every 5 minutes
2 measurements per beat
Readings kept for 14 days
1 callback endpoint
Get started
Builder
$39
per month
50 Signals · 50 Monitors · 2.16 million pooled stored events per 30 days
As often as every minute
8 measurements per beat
Readings kept for 90 days
5 callback endpoints
Get started
Smart
$199
per month
200 Signals · 200 Monitors · 8.64 million pooled stored events per 30 days
As often as every minute
16 measurements per beat
Readings kept for 365 days
20 callback endpoints
Get started
Scale
$999
per month
1,250 Signals · 1,250 Monitors · 54 million pooled stored events per 30 days
As often as every minute
16 measurements per beat
Readings kept for 365 days
50 callback endpoints
Get started
Additional monitors cost $0.50 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.
Each monitor slot includes one signal slot, and several monitors can share one signal. Signals have no separate charge.
Stored events are pooled across your organization; buying additional monitors does not increase the plan’s stored-event allowance.
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.
Unretained empty heartbeat receipts are not individually metered.
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?
Paid plans charge $2.50 per million stored events above the pooled allowance, within the plan cap and any lower customer cap. The plan cap equals the base monthly plan price. Free has no overage and refuses stored beats above its allowance.
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! Publish the count to a Signal. A linked Monitor can check it against a fixed minimum, such as 1, on its schedule.
Can it watch business numbers, not just jobs?
Yes! Publish orders, signups or refunds as numeric readings to a Signal, then link a Monitor with fixed numeric limits.
Can several Monitors use one Signal?
Yes. Several Monitors can link to one Signal and share its readings. Each Monitor has its own fixed-threshold numeric rules, schedule, incidents and callback.
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 I publish before creating a Monitor?
Yes. Create a Signal before any Monitor. A first publish with a manage key may create the Signal, never a Monitor. The Signal owns the publish URL, readings, explicitly chosen heartbeat period (or disabled), and missing-beat callback. Link one or more Monitors afterward.
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
Create a Signal. Then send your first reading.
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 →