Only real outages reach you.

StatusTick checks your product and reads every failed check before it pages you. Nothing to host: your coding agent sets it up from your repo.

Checkout API · Last nightExample

23:51Deploy
01:15Slow
02:30Bot wall
03:12Woke you
04:30–05:00Maintenance

Start from your code.

No monitoring server to run and no config to write. Your agent reads the repo and proposes the monitors. Nothing is created until you say yes.

Read https://statustick.com/setup.md and follow it.

Works with Claude Code, Cursor, VS Code, Codex CLI, Windsurf, Gemini CLI and Claude Desktop. Set up with an AI agent

Your agent, in ~/checkout-app Example

I found 7 things to monitor.

Monitors the agent plans to create
Type	Target	How often
HTTP	https://app.example.com	every 60 s
HTTP	/api/health	every 60 s
HTTP	/login · sign-in	every 60 s
HTTP	/checkout · checkout	every 60 s
HTTP	/api/v1 · main API	every 60 s
Heartbeat	nightly-report (vercel.json cron)	every 24 h
SMTP	smtp.example.com:587 · email	every 5 min

Private services, checked from inside.

The database, the internal API and the admin panel aren’t on the public internet. Run one small agent inside your network and they get the same checks and the same alerts as everything else.

Monitors Example

Checkout APIPublicHTTPS · 60 s
orders-dbPrivatePostgreSQL · SELECT 1
admin.corp.examplePrivateHTTPS · 60 s
search.internalPrivateHTTP · 1,840 ms
sessions-cachePrivateRedis · PING

$ docker run -d --restart unless-stopped \
    -e STATUSTICK_TOKEN="sta_live_…" \
    ghcr.io/statustick/agent:1

Shortened. Runs as a Docker container, a Linux binary or a Helm chart. Private agents docs

Connects out, never in.
Each agent connects out to agent.statustick.com:443 with its own token. You open no ports.
Your rules, on your machine.
An allow list on the agent decides what it may check, and nobody at StatusTick can change it.
Databases, read-only.
It logs in, runs one read-only command, and only the one value you compare leaves your network.
Keeps checking offline.
Results are kept for up to an hour and sent when the connection is back.

Every failed check is read before anyone is woken.

You choose how many failed checks in a row open an incident. The rules are written down, and you can see and change every one.

Bot walls, slow answers and planned work don’t page you. A real outage does, and the alert says what broke. A plain 403 from your own app still counts as down.

How StatusTick avoids false alarms

What StatusTick does with each kind of answer
What your server said	What it means for you	What happens
HTTP 503 · 2 checks in a row	Checkout is down for your customers.	Incident opened, on-call paged
HTTP 503 · then 200	A restart, probably a deploy. Already back.	Noted on the timeline
HTTP 403 · bot challenge page	Your firewall is blocking our check, not your users.	One notice with the steps to allow us
HTTP 200 · 4,210 ms	Up, but slow enough that people notice.	Shown as Degraded
Any answer · maintenance window	Your planned work, doing what you planned.	Stays quiet
HTTP 503 · Stripe reports an incident	Down, and the cause is upstream.	Paged, with Stripe’s incident attached

Alert, incident and status page, one record.

Your team and your users see the same facts, whether the alert came from StatusTick or from a tool you already run.

Incident timeline Example

14:05:08MonitorCheckout API down: 503 on 2 checks in a row.Checked from outside, and from inside your network with a private agent.
14:05:09IncidentsIncident opened. You are paged in Slack.Acknowledge from the alert. Escalation if nobody does.
14:06:10Statusstatus.example.com shows Checkout as down, with a drafted update.Components follow your monitors, on your own domain.
14:06:12VendorsStripe reports no incident, so the bug is yours.Official status of vendors you depend on, on your incident timeline.
14:09:30IncidentsSentry: error rate on /checkout above 5%. Opened here too.Alerts from your other tools get the same acknowledge, escalation and alerts.
14:21:00MonitorCheckout answers 200 again. Resolved after 16 minutes.The timeline stays on the incident.

#ops-alerts · 14:05:09Example

Checkout API is down

Down since 14:05 UTC: 503 on 2 checks in a row.

Example Inc. status status.example.com

Checkout is down. We are looking into failed payments since 14:05 UTC.

WebsiteOperational

17 Jul: up
18 Jul: up
19 Jul: up
20 Jul: up
21 Jul: up
22 Jul: up
23 Jul: up
24 Jul: up
25 Jul: up
26 Jul: up
27 Jul: up
28 Jul: up
29 Jul: up
30 Jul: up
31 Jul: up
1 Aug: up
2 Aug: up
3 Aug: up
4 Aug: up
5 Aug: up
6 Aug: up
7 Aug: up
8 Aug: up
9 Aug: up
10 Aug: up
11 Aug: up
12 Aug: up
13 Aug: up
14 Aug: up
15 Aug: up
16 Aug: up
17 Aug: up
18 Aug: up
19 Aug: up
20 Aug: up
21 Aug: up
22 Aug: up
23 Aug: up
24 Aug: up
25 Aug: up
26 Aug: up
27 Aug: up
28 Aug: up
29 Aug: up
30 Aug: up
31 Aug: up
1 Sep: up
2 Sep: up
3 Sep: up
4 Sep: up
5 Sep: up
6 Sep: up
7 Sep: up
8 Sep: up
9 Sep: up
10 Sep: up
11 Sep: up
12 Sep: up
13 Sep: up
14 Sep: up
15 Sep: up
16 Sep: degraded
17 Sep: up
18 Sep: up
19 Sep: up
20 Sep: up
21 Sep: up
22 Sep: up
23 Sep: up
24 Sep: up
25 Sep: up
26 Sep: up
27 Sep: up
28 Sep: up
29 Sep: up
30 Sep: up
1 Oct: up
2 Oct: up
3 Oct: up
4 Oct: up
5 Oct: up
6 Oct: up
7 Oct: up
8 Oct: up
9 Oct: up
10 Oct: up
11 Oct: up
12 Oct: up
13 Oct: up
14 Oct: up

CheckoutDown

17 Jul: up
18 Jul: up
19 Jul: up
20 Jul: up
21 Jul: up
22 Jul: up
23 Jul: up
24 Jul: up
25 Jul: up
26 Jul: up
27 Jul: up
28 Jul: up
29 Jul: up
30 Jul: up
31 Jul: up
1 Aug: up
2 Aug: up
3 Aug: up
4 Aug: up
5 Aug: up
6 Aug: up
7 Aug: up
8 Aug: up
9 Aug: up
10 Aug: up
11 Aug: up
12 Aug: up
13 Aug: up
14 Aug: up
15 Aug: up
16 Aug: up
17 Aug: up
18 Aug: up
19 Aug: up
20 Aug: up
21 Aug: up
22 Aug: up
23 Aug: up
24 Aug: up
25 Aug: up
26 Aug: down
27 Aug: up
28 Aug: up
29 Aug: up
30 Aug: up
31 Aug: up
1 Sep: up
2 Sep: up
3 Sep: up
4 Sep: up
5 Sep: up
6 Sep: up
7 Sep: up
8 Sep: up
9 Sep: up
10 Sep: up
11 Sep: up
12 Sep: up
13 Sep: up
14 Sep: up
15 Sep: up
16 Sep: up
17 Sep: up
18 Sep: up
19 Sep: up
20 Sep: up
21 Sep: up
22 Sep: up
23 Sep: up
24 Sep: up
25 Sep: up
26 Sep: up
27 Sep: up
28 Sep: up
29 Sep: up
30 Sep: up
1 Oct: up
2 Oct: up
3 Oct: up
4 Oct: up
5 Oct: up
6 Oct: up
7 Oct: up
8 Oct: up
9 Oct: up
10 Oct: up
11 Oct: up
12 Oct: up
13 Oct: up
14 Oct: down

Updates

Investigating · 14:06 UTC

Payments at checkout are failing. Next update within 30 minutes.

Emailed to subscribers

Alerts from Grafana, Prometheus Alertmanager, Datadog, Sentry, CloudWatch or any JSON webhook open incidents here too.
Alerts go out to email, Slack, Microsoft Teams, Discord, Telegram, PagerDuty or your own webhook.
Vendor status comes from the official pages of the services you follow. See the vendors we follow

What it checks.

Check types and where they run
Check	What it is for	From outside	With a private agent
HTTP(S)Websites, APIs and health routes. Checks the answer, not just that one came back.	Websites, APIs and health routes. Checks the answer, not just that one came back.
HeartbeatCron jobs and backups that stop reporting in.	Cron jobs and backups that stop reporting in.		through the relay
PlaywrightA real browser runs your sign-in or checkout.	A real browser runs your sign-in or checkout.
SSL and domainsCertificates and domains before they expire.	Certificates and domains before they expire.
DNS, TCP and pingRecords, open ports and round-trip time.	Records, open ports and round-trip time.
SMTP and IMAPYour mail servers greet like they should.	Your mail servers greet like they should.
MCP and gRPCYour MCP server and gRPC health service answer.	Your MCP server and gRPC health service answer.
PostgreSQL and MySQLLogs in and runs a read-only query.	Logs in and runs a read-only query.	—
Redis and MongoDBRuns the health command.	Runs the health command.	—
KubernetesAnnotated Services become monitors on their own.	Annotated Services become monitors on their own.	—

Questions.

What does it cost?: Nothing while in beta. Pricing comes with launch.
Do I need to host anything?: No. StatusTick runs the checks. A private agent is optional, and only for services that aren’t on the public internet.
What can my coding agent do with it?: It adds the StatusTick MCP server, plans the monitors with you and creates them after you say yes. The server cannot delete anything.
Do I have to replace my other tools?: No. Alerts from Grafana, Prometheus Alertmanager, Datadog, Sentry, CloudWatch or any JSON webhook open StatusTick incidents, with the same paging and status page.
Where do alerts go?: Email, Slack, Microsoft Teams, Discord, Telegram, PagerDuty or your own webhook.
When can I use it?: The closed beta is open now, and StatusTick launches in Q1 2027. We write when your invite is ready.

See it on your own product.

Point it at your product tonight. In the morning you’ll see what happened while you slept, and which of it needed you.

Only real outages reach you.

Start from your code.

Private services, checked from inside.

Connects out, never in.

Your rules, on your machine.

Databases, read-only.

Keeps checking offline.

Every failed check is read before anyone is woken.

Alert, incident and status page, one record.

What it checks.

Questions.

See it on your own product.