Uptime Monitoring: What It Is and How to Set It Up

Farouk Ben. - Founder at OdownFarouk Ben.()
Uptime Monitoring: What It Is and How to Set It Up - Odown - uptime monitoring and status page

Uptime monitoring is the practice of checking a website, API or server from the outside at regular intervals and alerting you when it stops responding correctly. A monitoring service sends a request to your URL every minute or few minutes, checks that the answer is what it should be, and notifies you by email, SMS, Slack or another channel when it isn't. It also records the results, so you can see your uptime percentage and response times over time.

The point is simple: you should hear about an outage from your monitoring, not from your customers.

This guide covers how uptime monitoring works, what to monitor, how to set it up so the alerts are worth reading, and what to look for in a tool.

How uptime monitoring works

Every uptime monitor follows the same loop:

  1. Send a request. The service makes a request to your URL from one of its servers, usually an HTTP(S) request just like a browser or API client would make.
  2. Check the response. It looks at whether a response came back at all, how long it took, the HTTP status code and, optionally, the content of the response.
  3. Decide whether it's down. A single failed request can be a network blip, so good tools retry and wait for several failed checks in a row before declaring an outage.
  4. Alert. Once the outage is confirmed, it notifies the people or channels you chose. When the site responds correctly again, it sends a recovery alert.
  5. Record. Every check is stored, which gives you uptime percentages, response time history and a record of each outage.

Because the checks come from outside your infrastructure, uptime monitoring catches the problems your users actually see: DNS failures, expired certificates, a crashed web server, a load balancer returning errors, or a hosting provider outage. Internal metrics can look healthy while all of these are happening.

Uptime monitoring vs other kinds of monitoring

Uptime monitoring answers one question: can people reach this, and does it respond correctly? It complements, rather than replaces, other tools:

  • Application performance monitoring (APM) looks inside your code at slow database queries and errors.
  • Infrastructure monitoring tracks CPU, memory and disk on your servers.
  • Real user monitoring (RUM) measures what real visitors experience in their browsers.

Many teams start with uptime monitoring because it is the quickest to set up and catches the most visible failures.

What uptime percentages mean

Uptime is the share of time a service was available. The differences between "nines" look small but add up:

Uptime Downtime per year Downtime per month
99% 3.65 days 7.3 hours
99.9% 8.76 hours 43.8 minutes
99.95% 4.38 hours 21.9 minutes
99.99% 52.6 minutes 4.4 minutes

If you promise customers a level of availability in an SLA, uptime monitoring is how you measure whether you met it. Our SLA calculator converts any percentage into allowed downtime, and the guide to five nines availability explains what it takes to reach the top of the table.

What to monitor

Monitoring only your home page is the most common mistake. A home page can load from a cache while sign-up, checkout or the API behind your app is failing. Start with:

  • Your home page and main landing pages, the pages new visitors and search traffic arrive on.
  • Login and sign-up pages, because a broken login locks out every existing customer.
  • Checkout or payment pages if you sell online.
  • API endpoints, especially a health-check endpoint that confirms the API and its database are working. See API monitoring for what a good health check returns.
  • Third-party services you depend on, such as a payment provider or authentication service, if they offer a URL you can check.
  • Subdomains such as app., api., docs. and your status page, which often run on different infrastructure from the main site.

How to set up uptime monitoring

1. Choose what counts as "up"

A page that returns a 200 status code isn't always working: an error page or an empty response can still return 200. Decide what a healthy response looks like:

  • Status code. Usually any 2xx response. For an API you might require an exact code, such as 200 or 204.
  • Response content. For a health-check endpoint, check that the body matches what a healthy service returns, for example {"status":"ok"}.
  • Method, headers and body. APIs often need an Authorization header or a POST request with a JSON body. Use a dedicated read-only token for monitoring.

2. Pick a check frequency

Check frequency sets how quickly you can find out about an outage. If a tool checks every minute and confirms an outage after three failed checks, you'll hear about it roughly three minutes after it starts. With five-minute checks, the same rule means about fifteen minutes.

For anything customer-facing, check every minute. Internal tools and low-traffic pages can be checked every 5 to 15 minutes.

3. Check from more than one location

A site can be down in one region and fine in another because of a CDN problem, a routing issue or a regional cloud outage. Checking from several locations tells you whether a problem is global or local, and the response times from each region show where users are getting a slow experience. Global uptime monitoring goes into this in more detail.

4. Avoid false alarms

Alerts that turn out to be nothing teach people to ignore alerts. To keep them meaningful:

  • Retry before failing. One timed-out request shouldn't count as an outage.
  • Require consecutive failures. Confirm an outage only after two or three failed checks in a row, and confirm recovery the same way.
  • Set realistic timeouts. Slow is not the same as down; record response times and investigate trends separately.

Our guide to false positives in uptime monitoring covers the common causes and fixes.

5. Send alerts where they'll be seen

Route alerts to channels your team actually watches: SMS or an on-call tool for urgent production outages, and Slack, Discord or email for everything else. If you have an on-call rotation, send alerts to PagerDuty or Opsgenie so escalation and schedules are handled there. Make sure more than one person gets alerts for critical services.

6. Watch your SSL certificates

An expired certificate makes a site unreachable for most visitors, and certificate lifetimes are getting shorter: public certificates issued from 15 March 2026 last at most 200 days, falling to 100 days in 2027 and 47 days in 2029. Choose a tool that warns you before certificates on your monitored sites expire.

7. Tell your users what's happening

When something does go down, a status page answers "is it just me?" before customers open support tickets. If the status page is linked to your monitors, it updates automatically during an outage.

What to look for in an uptime monitoring tool

When comparing uptime monitoring services, check:

  • Check frequency on the plan you'd actually buy, not just the most expensive one.
  • Number and spread of locations, and whether you can choose them.
  • Request options: HTTP methods, custom headers, request bodies and expected status codes, which matter as soon as you monitor an API.
  • False alarm protection: retries and confirmation after several failed checks.
  • Alert channels that match how your team works, such as email, SMS, Slack, Discord, Telegram, webhooks, PagerDuty and Opsgenie.
  • SSL certificate expiry alerts.
  • Status pages, ideally on your own domain and updated from your monitors.
  • Monitor limits and price as you grow.

Odown checks websites and APIs from up to 17 locations as often as every minute. You can set the HTTP method, headers, body, expected status code and expected response, and choose how many failed checks confirm an outage (three by default). Alerts go to email, SMS, Slack, Discord, Telegram, webhooks, PagerDuty or Opsgenie. The certificate of every monitored site is checked daily, and status pages on your own domain are included. Plans start at $12 a month with a 14-day free trial. To test a single URL right now, try the free website uptime checker.

Frequently asked questions

What is the difference between uptime and availability?

They are often used interchangeably. Strictly, uptime is the time a system is running, while availability is the time it is usable by the people who need it. A server can be up while the website on it returns errors, which is why uptime monitoring checks the response rather than just whether the server answers.

How often should I check my website?

Every minute for anything customers rely on. The check frequency, multiplied by the number of failed checks needed to confirm an outage, is roughly how long it takes to find out about a problem.

Can uptime monitoring check APIs as well as websites?

Yes. An uptime monitor that lets you set the method, headers and request body can check API endpoints the same way it checks web pages, and can confirm the response contains what you expect. See API monitoring.

Does uptime monitoring slow down my site?

No. A check is a single request every minute or few minutes from each location, which is negligible next to normal traffic. For APIs, point monitors at lightweight health-check endpoints rather than expensive operations.

How much does downtime cost?

It depends on your business, but lost sales, support load, SLA credits and damage to trust all add up quickly. Our article on the cost of downtime explains how to estimate it for your own service.