Store
The checkout URL
The homepage answers 200 while the payment page returns 500. Monitor the deep URL, and add a keyword check where specific text must appear.
Uptime monitoring
Uptime monitor for websites, servers and URLs
HostTracker's uptime monitor requests your website, server or any URL from 300+ checkpoints in 158 cities, as often as every minute, and tells you when it stops answering. Before anything is called an outage, the failed check is re-run from up to seven other locations.
30 days, 100 monitors, 1-minute checks, no credit card. The address you enter is carried into your new account.
- Trusted since 2004
- 500,000+ websites monitored
- 300+ checkpoints worldwide
How a failed check becomes a confirmed outage and an alert
As often as every minuteHTTP, ping and port checks from 300+ checkpoints in 158 cities. You choose the regions for each monitor.
Confirmed before it alertsA failed check is re-run from up to seven other locations, and the majority decides whether it is an outage.
Uptime you can showUptime reports by email, daily to yearly by plan, and a public status page on every plan.
What an uptime check sends, and what counts as down
Each check is a real request from a real network, the kind a visitor's browser or an API client makes.
HTTP MONITOR DEFAULTS
method GET (HEAD, POST, PUT, PATCH, DELETE, OPTIONS) timeout 40 s, adjustable up to 100 s redirects followed, up to 20 hops down when the connection fails, the request times out, or the status is 4xx or 5xx
URL monitoring, not only the homepage
An HTTP monitor requests the exact URL, path and query string included, with your headers, body and basic-auth credentials. An expected status, such as 401 on a private page, can be excluded.
Server uptime by ping or port
A ping monitor sends four ICMP packets and passes when most come back. A port monitor opens a TCP connection, or sends UDP, to the port you name.
From every minute to once a day
Intervals run from 1 minute to 24 hours; a new monitor starts at 3 minutes. The shortest interval per plan
Confirmation
How the multi-location quorum confirms an outage
One checkpoint can fail for reasons unrelated to your site, such as a routing problem near its data center. So by default HostTracker does not alert on one result. When a check disagrees with the last confirmed state, this runs in order.
- One checkpoint sees a change. The scheduled check runs from one location in your chosen regions. A result that matches the confirmed state ends there.
- A 10-second pause, so the target is not hit again in the instant it changed.
- Up to seven locations re-run the check, drawn from the monitor's own pool, locations with a clean record first, started 2 seconds apart. If fewer than three such locations are available, others top the group up to three.
- Only countable results vote. A location that could not run the check is left out. One whose failure was its own, such as a resolver timeout or an overloaded agent, is left out, trusted less for a while and replaced.
- The majority decides. A tie keeps the previous state. If no location returns a countable result, the check runs again later instead of guessing.
- Then the alert. A confirmed Down opens an incident and starts the alerts. Recovery is confirmed the same way.
Timing. The last re-check starts 22 seconds after the first failure: the 10-second pause plus six 2-second steps. A site that times out adds up to its timeout on top.
Other rules. Per monitor, instead of the majority you can require a set number of locations to report Down, require all re-check locations to agree, require full agreement for Down with a majority for Up, or turn the re-check off. Multi-location availability monitoring covers the location roster in depth.
Uptime calculator
What an uptime percentage allows: the SLA table
An uptime commitment is easier to judge as time: the period length multiplied by the share the service may be down.
allowed downtime = period length x (100 - uptime %) / 100
The table uses a 30.44-day month (2,630,016 seconds) and a 365.25-day year (31,557,600 seconds), the averages SLA tables use and the ones the uptime calculator uses. Example: 99.9% of a month is 2,630,016 x 0.1 / 100 = 2,630 seconds, or 43 minutes 50 seconds. Values are rounded down to the second.
| Uptime target | Downtime per day | Per month | Per year |
|---|---|---|---|
| 99% | 14m 24s | 7h 18m 20s | 3d 15h 39m |
| 99.9% | 1m 26s | 43m 50s | 8h 45m 57s |
| 99.95% | 43s | 21m 55s | 4h 22m 58s |
| 99.99% | 8s | 4m 23s | 52m 35s |
Other targets, or downtime back to a percentage: the uptime and SLA calculator. Measuring the number and turning a miss into a credit is covered in SLA monitoring and SLA credits.
Where an uptime check earns its keep
Six addresses worth a monitor of their own
Each of these can fail while the homepage still looks fine.
API
An endpoint with its own method
POST a body with an authorization header; any 4xx or 5xx is down. To judge the JSON itself, use an API monitor.
Server
A VPS or dedicated server
Ping shows whether the machine answers; a port monitor on 22 or 443 shows whether its service does.
Mail
A mail server
Port monitors on SMTP (25 or 587) and IMAP (993) catch a mail service that stopped accepting connections.
Agency
Client sites in one account
Each client site with its own alert contacts, and a monthly uptime report by email for the client.
Private
A page behind a login
Send basic-auth credentials, or exclude the 401 your login wall returns, so the check measures the server, not the lock.
After the verdict
Alerts, reports and quiet maintenance windows
What happens once an outage is confirmed, and what is kept afterwards.
On the channel each person reads
Email, SMS, voice call, Slack, Microsoft Teams, PagerDuty, Telegram, Discord, Viber, Mattermost, Pushover, Pushbullet, browser push and webhooks. Each contact has its own delay, from immediately to 24 hours of continuous downtime.
Uptime reports and a status page
Uptime reports by email, daily to yearly as the plan allows. A public status page, on every plan, shows the same monitors to your users.
Quiet during planned work
Down alerts are held during a scheduled maintenance window. A site still down when the window ends gets one Down alert then.
Pricing
Plans and the shortest check interval each allows
Read from the live plan catalog, the same numbers as the pricing page. Prices are per month on monthly billing.
| Plan | Price | Monitors | Shortest interval | Status pages | Uptime reports |
|---|---|---|---|---|---|
| Complete | $9 a month | 1 website, all monitoring types | 1 minute | 1 | Daily to yearly |
| Webmaster | $14 a month | 10 | 5 minutes | 2 | Daily to yearly |
| Business | $29 a month | 25 | 1 minute | 5 | Daily to yearly |
| Enterprise | $99 a month | 150 | 1 minute | 10 | Daily to yearly |
Complete covers one website: all of its monitors belong to that site and its subdomains (fair use).
- 30-day money-back on your first payment
- Your price stays while your subscription stays active
Other billing periods and plan sizes are on the pricing page. The 30-day trial includes 100 monitors and 1-minute checks.
Moving from another uptime monitor?
The importer in the New monitor panel reads exports from UptimeRobot, Pingdom, StatusCake, Better Stack, Site24x7 and Uptime Kuma. It converts HTTP, keyword, ping and port monitors and lists anything it cannot convert, with the reason. A plain list of URLs, one per line, works too. With 100 monitors in the trial you can run both services side by side for a month.
Import your monitorsStart the 30-day trial first and confirm your email; the importer opens once the address is confirmed. HostTracker is not affiliated with any company named here; product names are trademarks of their owners.
Every layer of your stack, monitored
Websites, servers, APIs, certificates. One check type per page, the same locations, alerts and reports behind all of them.
Trusted by teams at
The full guide
Running an uptime monitor well
Every chapter opens in place, so the page stays short.
Choosing a check interval
The interval is the longest an outage can run before a check sees it; confirmation adds the re-check on top. A shorter interval also records short outages a longer one misses. A practical split: 1 minute for checkout, login and APIs, 5 to 15 minutes for marketing pages, 30 minutes or more for staging.
Choosing check locations
A new monitor checks from Western Europe, Eastern Europe and North America. Pick the regions your users are in: there are groups for regions and single countries, and a group can be subtracted, for example everywhere except one country whose routing should not decide your alerts. The re-check draws from the same pool, so a very narrow pool has fewer locations to confirm from.
Alert delays and recovery messages
A team chat at 0 minutes and an on-call phone at 5 minutes means a two-minute outage reaches the chat and wakes nobody. On recovery, the message goes only to contacts whose delay had passed: a contact never told about the outage is not told it ended. Repeated alerts while a site stays down are a separate subscription per contact.
Setting up an uptime monitor
- Run the instant HTTP check on the URL to see what a check gets back now.
- Add an HTTP monitor for the URL, or a ping or port monitor for a server, and set the interval.
- Pick the location groups; keep the majority re-check unless you have a reason not to.
- Subscribe contacts and set each one's delay.
- Add an uptime report for an email contact, and a status page if users should see the state.
Key takeaways
Key takeaways
- HostTracker's uptime monitor checks a website, server or URL from 300+ checkpoints in 158 cities, at intervals from 1 minute to 24 hours depending on the plan.
- By default a failed check is re-run 10 seconds later from up to seven other locations started 2 seconds apart, and the majority of countable results decides.
- An HTTP monitor counts a failed connection, a timeout (40 seconds by default) or a 4xx/5xx status as down; ping and port monitors cover servers.
- Alerts go by email, SMS, voice call, Slack, Microsoft Teams, PagerDuty, Telegram, webhooks and more, each contact with its own delay from immediately to 24 hours.
- 99.9% uptime allows 43 minutes 50 seconds of downtime in an average month; the 30-day trial includes 100 monitors and 1-minute checks without a credit card.
Frequently Asked Questions
A service that requests your website, server or URL on a schedule from outside your network, records whether and how fast it answered, and alerts you when it stops. Those records become your measured uptime. HostTracker checks from 300+ checkpoints in 158 cities, so the result reflects what users elsewhere see.
As often as every minute, depending on the plan, and as rarely as once a day. A new monitor starts at 3 minutes and the interval can be changed at any time. The pricing table on this page lists each plan's shortest interval; the 30-day trial allows 1 minute.
By default it does not alert on one result. A check that disagrees with the confirmed state is re-run 10 seconds later from up to seven other locations, started 2 seconds apart. Results a location could not produce, or that failed for the location's own reasons, do not count. The majority of the rest decides, and a tie keeps the previous state.
Yes. A ping monitor sends four ICMP packets and passes when most come back: does the machine answer. A port monitor connects to a port such as 22, 25 or a database port: does the service accept connections. One of each separates a dead machine from a dead service.
It allows 43 minutes 50 seconds of downtime in an average month and 8 hours 45 minutes 57 seconds in a year. A 30.44-day month is 2,630,016 seconds, and 0.1% of that is 2,630 seconds. At 99.99% the monthly allowance is 4 minutes 23 seconds.
Email, SMS, voice call, Slack, Microsoft Teams, PagerDuty, Telegram, Discord, Viber, Mattermost, Pushover, Pushbullet, browser push and webhooks. Each contact has its own delay: immediately, after 3, 5, 15, 30 or 60 minutes, or after 3, 6, 12 or 24 hours of continuous downtime.
Yes. The importer in the New monitor panel reads exports from UptimeRobot, Pingdom, StatusCake, Better Stack, Site24x7 and Uptime Kuma, and lists anything it cannot convert with the reason. A plain list of URLs works too. HostTracker is not affiliated with these companies; their names are trademarks of their owners.
Up to 100 monitors with 1-minute checks for 30 days, and no credit card. If you then choose a plan: 30-day money-back on your first payment; your price stays while your subscription stays active.
30-day trial, 100 monitors, no credit card
An uptime monitor that confirms an outage before it wakes anyone
Checks from 300+ checkpoints in 158 cities, every failure confirmed by a majority of up to seven locations, alerts on the channel each person reads.
- Trusted since 2004
- 500,000+ websites monitored
- 300+ checkpoints worldwide
Part of HostTracker's website monitoring software.
30-day trial - no card needed
Start the 30-day trial