Website monitoring from 22+ locations

Your site is up. Is it fast in Mumbai?

Checkly measures your website from 22+ locations around the world: availability, response time by phase, and whether the page actually works once it loads. A slow region shows up as a degraded alert, not as a global average that looks fine.

Start for freeTalk to an engineer

Free plan · 10 monitors · No credit card

GET https://www.acme.com/p95 · by location · last 24h
  1. N. Virginiaus-east-1160 ms
  2. Londoneu-west-2194 ms
  3. Frankfurteu-central-1216 ms
  4. São Paulosa-east-1340 ms
  5. Mumbaiap-south-11,041 ms · degraded
  6. Tokyoap-northeast-1342 ms
  7. Sydneyap-southeast-2412 ms
DNSTCPFirst byteDownloaddegraded > 800 ms · failed > 3,000 ms

Teams that watch their sites from the outside in with Checkly

Vercel
Carhartt
CrowdStrike
Airbus
Fanatics
Mistral
ServiceNow
GoFundMe
Hopper
1Password
Fastly
Total Wine
“A real advantage that Checkly gives us is that we’re not waiting for our users to report an issue with a product, or waiting on a staff member to file a ticket. Checkly gives us real-time feedback on what is and isn’t working.”
Thomas ReitherAssociate Consultant, Solutia

“The site is up” answers one question. Customers ask three.

A green uptime badge can hide a four-second first byte in one region and a checkout button that never enables. Checkly answers all three from the same platform, with the same alerting.

availability

Is it up, from where it matters

HTTP checks from 22+ locations, as often as every 10 seconds, with SSL expiry, status, and multi-region confirmation before anyone is paged.

response time

Is it slow, and in which phase

DNS, TCP, first byte, and download timing per region, with P50, P95, and P99 trends. A degraded threshold turns a slow region into a signal before it becomes an outage.

real browser

Does it work after it loads

Playwright checks that render the page, click through login and checkout, and fail when the confirmation never appears. A 200 with a broken script error is still a failure.

Alert on slow before you alert on down

Every monitor carries a degraded threshold and a failed threshold. Cross the first and the region goes amber with its own alert; cross the second and on-call is paged. Timing is split into DNS, TCP, first byte, and download so you can see which layer moved, per location, not as one misleading average.

  • P50, P95, and P99 response time per location, 30 days or more of history
  • Degraded and failed thresholds per monitor, with separate alerting
  • Multi-region confirmation and retries so a blip never pages anyone
storefront.check.tsts
1import { UrlMonitor, UrlAssertionBuilder, Frequency }
2 from 'checkly/constructs'
3import { slackWeb, pagerduty } from './alert-channels'
4
5new UrlMonitor('storefront', {
6 name: 'Storefront · home',
7 frequency: Frequency.EVERY_1M,
8 locations: [
9 'us-east-1', 'eu-west-2', 'eu-central-1',
10 'sa-east-1', 'ap-south-1', 'ap-southeast-2',
11 ],
12 request: {
13 url: 'https://www.acme.com/',
14 assertions: [UrlAssertionBuilder.statusCode().equals(200)],
15 },
16 // slow is a signal before down is an incident
17 degradedResponseTime: 800,
18 maxResponseTime: 3000,
19 alertChannels: [slackWeb, pagerduty],
20})
Checkly browser check result showing Playwright steps, timings, and a failed assertion

Prove the page works, not just that it loads

A browser check runs Playwright against your live site on a schedule: render the page, log in, search, add to cart, pay. It fails when the confirmation never appears, when a third-party script blocks the button, or when an API behind the page returns garbage with a 200.

  • Login, signup, search, and checkout flows in Chromium, Firefox, and WebKit
  • Web Vitals and step timings on every run
  • Screenshots, video, and the full Playwright trace kept with each failure
More on browser and API checks

Regressions arrive as evidence, not as a ticket

When a check fails, Checkly keeps the timing breakdown, screenshots, network calls, and console logs, then Rocky AI reads them and says what broke and where. The alert reaches Slack, PagerDuty, Opsgenie, Teams, SMS, phone, or any webhook with the evidence attached.

  • Root cause and user impact written for the responder
  • Per-location results, so a regional CDN or DNS problem is obvious
  • Public or private status pages fed by the same monitors
checkout-flow
Sydney · ap-southeast-2
✦Rocky AI · Root cause

The Buy now button never became clickable. The payments API returned 503 from ap-southeast-2, so the cart never advanced to confirmation. Users in that region could not complete a purchase.

Evidence2
  • •

    Step timed out on getByRole('button', { name: 'Buy now' }) after 10000ms.

    trace
  • •

    POST /api/payments returned 503 Service Unavailable, 4 consecutive runs.

    network
“Checkly is incredible: It combines Pingdom, Ghost Inspector and Assertible all in the same app, and the insights are much more detailed.”
Leo LamprechtSVP Product, Vercel

Frequently asked questions

Straight answers on response-time metrics, locations, slow-versus-down alerting, RUM, and the free plan.

Website monitoring checks a site from outside your infrastructure on a schedule, from locations around the world, and records whether it responded, how long each phase took, and whether the page actually worked once it loaded. Checkly combines URL monitors for availability and response time with real-browser checks that walk through the page the way a customer would.

URL monitors record availability and response time per location, broken down into DNS, TCP, first-byte, and download phases, with P50, P95, and P99 trends over time. Browser checks add Playwright step timings, Web Vitals, screenshots, and traces for the rendered page.

Yes. Checkly runs checks from 22+ public locations across North America, South America, Europe, Asia, Australia, and Africa, so you see performance the way customers in each region experience it. Private locations run the same checks from inside your own network.

Yes. Every URL monitor has a degraded response-time threshold and a failed threshold. Cross the degraded line and the monitor reports degraded with its own alert, so a slow region surfaces while the site is still technically up.

Real user monitoring reports what visitors already experienced. Synthetic monitoring runs controlled checks on a schedule, including at 3 a.m. with no traffic, from locations you choose, so you find the slow or broken page before a customer does. Most teams run both.

No. Create URL monitors and browser checks in the web app with no code. Teams that prefer it define every monitor as a TypeScript construct in the Checkly CLI, versioned in git and deployed from CI, with Terraform and Pulumi providers too.

The free Hobby plan includes 10 uptime monitors across 6 locations plus browser and API check runs, with no credit card required. Paid plans add more monitors, faster intervals, all 22 locations, and private locations. See the pricing page for the full breakdown.

See your site the way Sydney sees it

Paste a URL, pick your locations, set a degraded threshold. The next slow region shows up on your screen before it shows up in support.

Start for freeTalk to an engineer

Free plan includes 10 monitors and 6 locations. No credit card.