Downtime Alerts by Phone Call: Website Down Alerts That Reach You in 60 Seconds
Email is easy to sleep through. TrackDowntime checks your site every 60 seconds from outside your hosting, then calls your phone, texts you and emails you, and keeps escalating until someone answers and the site is back up.
Phone call and SMS alerts start on Professional at $99 a month. Alerts are never billed per message.
Paste the URL, create your account, and the first check runs right away.
How to get a phone call when your website goes down
To get a phone call when your website goes down, point an external monitoring service at the URL, set the check interval to 60 seconds, and add your mobile number as a voice alert channel. When two consecutive checks fail, the service places an automated call. TrackDowntime does this on the Professional plan at $99 a month, and repeats the call until the site recovers.
Last updated September 2026
What a downtime alert has to do to actually be useful
Every monitoring tool can send an email. The difference between tools shows up at 3am on a Saturday, when the email lands in a inbox nobody is looking at and the checkout page stays broken until a customer complains.
An automated phone call, not just a notification
A push notification competes with every other notification on the phone. A ringing phone does not. When an endpoint fails its checks, TrackDowntime places an automated voice call that states which site is down, so the person who can fix it is awake and knows where to look before they open a laptop.
Escalation that repeats until recovery
One missed alert should not cost you the night. Alerts repeat while the endpoint is still failing instead of firing once and going quiet, so a call that goes to voicemail is followed by another attempt rather than silence.
A separate recovery notification
Knowing the site is back matters as much as knowing it broke. When the endpoint starts responding normally again, you get a distinct recovery message with the incident closed, so nobody drives to the office for an outage that already ended.
Checks from outside your own hosting
A monitor that runs on the same server as the site goes down with the site. TrackDowntime checks from infrastructure that has nothing to do with your host, which is the only way to catch a failure where the whole box, the DNS or the certificate is the problem.
No per message alert billing
Several monitoring tools sell SMS and voice as credit packs, so a bad week of flapping alerts turns into an invoice surprise or, worse, a silent stop when the credits run out. Alert volume here is included in the plan and does not meter.
Alerts on the endpoints that carry money
Monitor the checkout, the booking form, the login and the API, not only the homepage. A homepage can render perfectly while the payment endpoint returns a 500, and that is the failure that actually costs revenue.
Setting up phone call downtime alerts
Add the URL you cannot afford to lose
Start with the page that makes money: the checkout, the booking flow, the quote form or the API base URL. Paste the full URL including https. The first check runs immediately so you can confirm the monitor sees the endpoint as up.
Set the check interval to 60 seconds
Sixty second checks mean a failure is confirmed and alerted inside a couple of minutes rather than up to ten. Five minute checks on the Starter plan are fine for a brochure site, but anything transactional should be on the 60 second interval.
Add the phone number that gets the call
Enter the mobile number in full US format. This is the number the automated voice call reaches when an incident opens. Use the phone of whoever is actually able to act at 3am, which is often not the account owner.
Confirm the escalation and recovery messages
Check that the alert repeats while the incident is open and that recovery notification is on. Then leave it alone. A monitor is the one tool you want to forget about until the day it earns its entire year of cost back.
Why email alone keeps failing at the moment it matters
Downtime is not evenly distributed across the working day. Deploys land in the evening, certificates expire at midnight UTC, hosting providers do maintenance on weekends, and traffic spikes hit on the days you least want a failure. That means a large share of real outages start while nobody is watching a screen.
An email sent at 2:14am is not an alert. It is a record of an outage you will read about later, which is a different product. By the time it is opened, the site has often been down for hours, the ad spend for the morning has already run against a broken checkout, and the first person to notice was a customer.
This is the whole reason voice alerting exists. A phone call is the only channel most people have not trained themselves to ignore, because it is the channel reserved for emergencies and family. Using it for a failing production endpoint is the point.
How fast you find out, and why the interval matters more than it looks
Two numbers decide how long you are down before anyone knows: how often the endpoint is checked, and how many failed checks are required before an incident opens. Confirming a failure across consecutive checks is deliberate, because a single timeout is often a network blip rather than an outage, and a tool that calls you for every blip gets muted within a week.
On a 60 second interval, a genuine failure is confirmed and escalated within a few minutes. On a five minute interval, the same failure can run ten minutes or more before the first alert leaves. For a brochure site that difference is academic. For a store taking orders overnight, or a booking system where a missed slot is a lost customer, ten quiet minutes is the difference between an incident and a refund conversation.
The other half of the equation is where the check runs from. Checking from outside your own hosting is what catches the failures that matter most, because whole server failures, expired certificates and DNS problems are invisible to anything running inside the same environment.
Deciding who the call should actually go to
The most common setup mistake is routing the alert to the person who owns the billing rather than the person who owns the fix. If the call wakes someone who can only forward it, you have added a step to the outage rather than removing one.
For a solo operator this is simple, because it is you. For a small team, pick the person who can restart the service or roll back the deploy, and make sure a second number exists for the nights that person is unreachable. If you look after sites on behalf of other people, the alert should reach you before it reaches the client, which is the entire argument for website monitoring for agencies: the client should never be the one who tells you their site is down.
Worth checking before you commit to any tool: whether alert volume is billed per message. A flapping endpoint can fire dozens of alerts in a night, and on credit based pricing that either costs real money or stops silently when the balance empties. We break the line items down in our comparison of uptime monitoring pricing per site.
What to monitor beyond the homepage
Homepage only monitoring produces a false sense of safety, because the homepage is usually the most heavily cached and least fragile page on the site. It can serve a perfect 200 from cache while the parts of the application that touch a database are failing.
A practical set for a small commercial site is three to four endpoints: the homepage, the checkout or booking step, the login, and the API endpoint your mobile app or integrations depend on. That is well inside the five endpoints on the Starter plan for a single site, and it is why agencies and multi site operators tend to size at two to four endpoints per property rather than one. The full list of what is recorded on every check is on the monitoring features page.
Who sets up phone call alerts first
Stores taking orders overnight
Ecommerce traffic does not stop at 6pm, and neither does ad spend. If the checkout breaks at midnight and the alert is an email, the first signal is a support ticket in the morning plus a full night of paid traffic sent to a broken funnel.
Booking and appointment systems
A restaurant, clinic or salon booking page that is down during the evening rush loses bookings that never come back, because the customer books somewhere else within the minute rather than trying again later.
Solo operators and small teams with no on call rota
Most small businesses will never staff a night shift or run a formal on call rotation. A phone call is the cheapest possible substitute for one, and for a single person business it is the only realistic option.
APIs that other businesses depend on
If customers integrate against your API, an outage is a breach of an expectation and sometimes of a contract. Monitoring the API base URL separately from the marketing site is the minimum, because the two fail independently.
Freelancers and agencies looking after client sites
When you are paid to keep a site up, finding out from the client is a retainer problem before it is a technical one. Alerts that reach you first turn an awkward email into routine maintenance work.
Sites that just came back from an outage
The most common day to buy monitoring is the day after an outage nobody caught. If that was this week, start with the endpoint that failed and the one next to it in the funnel.
How downtime alerting compares across common tools
The table below is about alerting behaviour specifically, not overall feature depth. Every tool here is a legitimate choice and several are cheaper than we are. What differs is what happens after the first alert fires.
| Tool | Voice call alerts | Alert volume billing | Repeats while still down |
|---|---|---|---|
| TrackDowntime | Yes, on Professional and Enterprise | Included, not metered per message | Yes, until recovery |
| UptimeRobot | Yes, on paid plans | Free plan: SMS and voice bought separately. Paid plans include a monthly allowance | Not by default |
| StatusCake | Yes | Included, with rate limits on calls | Limited: documented as one call per incident |
Checked against each vendor's public pricing and documentation pages in September 2026. Monitoring vendors change plans and allowances often, so confirm the current terms before you buy. UptimeRobot publishes its pricing in euros.
Questions people ask before setting up downtime alerts
How do I get a phone call when my website goes down?
Add the URL to an external monitoring service, set the check interval to 60 seconds, and register your mobile number as a voice alert channel. When the endpoint fails consecutive checks, the service places an automated call telling you which site is down. On TrackDowntime this is included on the Professional plan at $99 a month.
How do I get notified when my website is down?
Pick the channels by how urgent the endpoint is. Email is enough for a brochure site, SMS is reasonable for a site that takes enquiries, and a phone call is the right choice for anything handling payments or bookings outside office hours. TrackDowntime sends all three and escalates rather than firing once.
How fast will I know that my site is down?
On a 60 second check interval, a genuine outage is confirmed across consecutive failed checks and escalated within a few minutes. On a five minute interval the same failure can run ten minutes or longer before the first alert is sent, which is why transactional endpoints belong on the shorter interval.
What happens if I miss the first downtime alert?
Alerts repeat while the incident is still open rather than firing once and stopping. A call that goes to voicemail is followed by further attempts, and the alert continues across channels until the endpoint responds normally again. A missed first alert does not mean silence for the rest of the night.
Do downtime alerts cost extra per message?
Not here. Phone call and SMS alerts are included in the Professional and Enterprise plans and are not metered per message. This matters when an endpoint flaps, because credit based pricing at other vendors can either generate an unexpected invoice or stop sending alerts once the balance runs out.
Can downtime alerts go to more than one person?
Yes. Alerting a second number is the standard way to cover holidays, flights and the nights when the primary contact is unreachable. For teams looking after client sites, the usual pattern is one internal number that always gets the call plus a backup.
Will I get an alert when the website comes back up?
Yes. Recovery is sent as a separate notification with the incident marked resolved, so you know the outage is genuinely over rather than assuming it from the absence of new alerts. The incident duration is recorded in the uptime history for that endpoint.
Can I monitor an API endpoint rather than a web page?
Yes. Any HTTP or HTTPS endpoint that returns a status code can be monitored, including API base URLs and health check routes. Monitoring the API separately from the marketing site is worth doing, because a static homepage and a database backed API fail for completely different reasons.
Put a phone number behind your most important URL
Add the endpoint that carries your revenue, set the interval to 60 seconds, and add the number that should ring. Starter is $29 a month for 5 endpoints with email alerts. Professional is $99 for 25 endpoints at 60 second checks with SMS and phone call alerts, and annual billing is roughly half the monthly rate.
Keep reading
Watching a portfolio of client sites from one account, endpoint sizing and per client cost.
Uptime monitoring pricing per siteWhat monitoring costs per endpoint across the main US tools, and the fees that inflate the bill.
Plans and pricingEndpoint counts, check intervals and which plans include SMS and phone call alerts.
Monitoring featuresWhat is recorded on every check, escalation behaviour and downtime history.