Get early access
Features · Monitoring · Uptime Monitoring
Monitoring · Uptime Monitoring

Outside-in checks,
where your customers are.

HTTP and HTTPS probes from multiple regions, attached automatically to every domain we discover. Root + path-level. Status, latency, body match, header match, certificate validity — one alert per real outage, not per check.

probes · 3 regions
acme.io 200 · 142ms
api.acme.io/health 200 · 88ms
shop.acme.io 200 · body match ✓
admin.acme.io 301 · redirect
cdn.acme.io cert expires 12d
— Why this matters

Inside metrics lie. Outside checks do not.

01

A healthy box can still be unreachable.

Firewall rule, load balancer config, DNS pointing somewhere else — none of it moves the agent's metrics.

02

One probe is one opinion.

A single failing check might be the probe's connectivity. Two regions failing is your problem.

03

Checks belong to the domain.

Every check lives under the domain it tests, with a primary check on the root and as many path checks as you need. No separate list to keep in sync.

— What's included

What a check can assert.

Status
Expected code, configurable. 3xx, 4xx and 5xx ranges raise events.
Latency
TTFB and total, p95 across regions, alertable.
Body match
Substring or regex — this catches the 200 that renders an error page.
Header match
Cache-Control, X-Vary and anything else you regressed.
Certificate validity
Expiry, chain and hostname match, with alerts at 30, 14 and 3 days.
Authenticated endpoints
Per-domain credentials, stored encrypted and never logged.
Sub-checks per path
Added from the domain with the prefix locked, so a typo cannot point a check at someone else's host.
Seven probes, seven networks
Across distinct autonomous systems, grouped by region — NA, EU, APAC, SA, MEA — so a regional outage reads as one.
Probes that recover themselves
A probe that drops out re-authenticates and comes back, and its return fires a recovery notification rather than staying silently absent.
— What it looks like

Response time, from outside.

180 ms timeout budget
Example view · https://acme.io · last 24 h · probes from eu-fra, us-iad and ap-sgp
— How it works

Outside in, from more than one place.

1

Probe from several regions

Each check runs from more than one location, because one probe is one opinion about whether you are up.

2

Assert more than a 200

Status, latency, a body match, a header match and certificate validity all count toward pass or fail.

3

Confirm before paging

A single failure is usually a network hiccup. Agreement between probes is what becomes an event.

— Works with

Connected capabilities

Domain Discovery
The list of domains that get monitored — built without you maintaining it.
→
Heartbeats
Counterpart for things that send a signal rather than respond to one.
→
Public Status Pages
Branded status pages built from these checks.
→
Notification Channels
Where the uptime alerts go.
→
— FAQ

Questions, answered honestly.

How many regions probe my domain? +
Two on Free, four on Starter, eight on Pro. Multi-region consensus suppresses single-probe noise.
Check frequency? +
60 s on Free, 30 s on Starter, 10 s on Pro. Custom intervals on Enterprise.
Can I monitor an internal domain? +
Yes — install the agent on a host inside the network and tell us to probe from that host. Same alerting, different vantage point.

Watch your endpoints, not just your boxes.

Free tier includes uptime checks on 5 domains.

Get early access Back to Monitoring