Hosted uptime monitoring
A monitor is an automated check that visits your website or service on a regular schedule and records whether it responded correctly. When the monitor lives on the same server as the thing it watches, a single failure takes both of them down at the same time, and the failure is the very moment when you most need an alert. DownForReal runs your monitors on separate infrastructure that we operate, so the alert can still reach you when your own server cannot.
The trial lasts for 14 days. You can import your existing monitors from UptimeRobot in about 2 minutes.
A failure domain is a set of systems that can fail together because they depend on the same server, the same network connection, or the same power supply. A self hosted monitor usually shares a failure domain with the service it watches. Self hosted means that you install and run the software yourself on your own machine. That arrangement works well on ordinary days, and it fails silently on the one day when the server, the network, or the building loses power. DownForReal removes that shared fate by placing the monitor, the alerting system, and your public status page on infrastructure that sits outside your failure domain. A status page is a public web page that shows visitors whether your services are working at the present moment.
How it works
The setup takes a few minutes, and the daily work of running servers, applying updates, and keeping the monitoring system itself online belongs to us after setup is complete.
You give us the address of your website, service, or server, and you choose how you want to be told when something goes wrong. You can receive alerts by email and through the chat and incident tools that your team already uses.
If you already use UptimeRobot, you can import your existing monitors instead of typing them again. The import process is described in detail further down this page.
Your instance is a separate, private copy of the monitoring software that runs for your account alone. That software is powered by Uptime Kuma, which is a widely used open source monitoring tool. Open source means that the source code is public, so anyone can read how the software works and check what it does.
We handle the hosting, the backups, and the software updates. You open a dashboard in your browser and you see the same clear view of uptime, response time, and recent incidents that self hosted users value.
When a check fails, our confirmation system checks the same address from two additional regions before any alert is sent. That confirmation step is explained in the next section, and it exists to protect your attention.
When all three regions agree that your service is down, we send the alert and we update your status page. Your status page stays reachable during your outage because it lives on our infrastructure rather than on yours.
Most customers begin with a homepage and a checkout or sign in flow, and then they add the supporting services that those pages depend on.
We check that the page loads, that it returns a healthy status code, and that it responds within a reasonable time.
An API, which is short for application programming interface, is a way for software programs to talk to each other. We check that your API answers requests and returns the response that you expect.
A port is a numbered entry point that a server uses for a particular kind of connection. We check that the port accepts connections, so you learn about a failure even when no web page is involved.
We can confirm that an expected word or phrase appears on a page, and we can warn you before a security certificate expires. A security certificate is the digital credential that lets a browser connect to your site over an encrypted connection.
Three region confirmation
A single failed check can have ordinary causes. A routing problem in one city, a brief network interruption, or a slow connection can make a healthy website look down from one location. An alert that fires on that single signal teaches you to ignore alerts, and ignored alerts are worse than no alerts at all.
One of our checking locations tries your address and receives no healthy response. At this point we record the failure, and we send no alert to you.
Checking locations in two other geographic regions test the same address right away. Each region reaches your service through a different network path.
If all three regions report the service as down, we treat the outage as confirmed and we notify you. If any region reaches your service successfully, we treat the first failure as a local network problem and we keep watching quietly.
This small demonstration uses the same rule that the service uses. Press the button, and the page will check the example address from all three regions in sequence.
This demonstration runs entirely in your browser. It illustrates the confirmation rule, and it performs no real check against any server.
Proof, in public
Dogfooding means that a company uses its own product for real work inside the company before it asks customers to rely on that product. In this service, dogfooding takes a specific and public form. The service monitors its own infrastructure in public, and anyone can read the resulting record.
The control plane is the central system that manages accounts, stores monitor settings, and coordinates every check. A probe region is one of the geographic locations that performs the checks. The word fleet means all of those systems taken together, including the control plane and the checking locations in all three regions.
The central system that runs accounts, dashboards, and alert delivery is monitored by the service itself.
The checking location in North America is monitored in public, along with the service that it supports.
The checking location in Europe is monitored in public, along with the service that it supports.
The checking location in Asia Pacific is monitored in public, along with the service that it supports.
The fleet uptime record is published as the trust record for the service. Uptime means the share of time that a system was available and responding correctly. Because the record is public, you can watch our availability over time, you can see incidents when they happen to us, and you can judge our reliability from evidence rather than from promises.
A real customer site, from day one
The service also monitors 5Quotes from day one. 5Quotes is a real customer site, and its monitors run through the same checks and the same three region confirmation process that every other customer receives.
5Quotes is presented here for one reason only. The site is a live site that the service monitors, and that fact gives you a concrete example of the service doing real work for a real site from the first day.
If the service can earn trust by watching itself in public and by watching a real customer site from day one, then the same system is ready to watch your site with the same care.
Moving from UptimeRobot
Switching monitoring tools usually means rebuilding every monitor by hand, and that busywork is the reason many teams stay with a service that no longer fits them. Our importer reads the monitor list that you export from UptimeRobot and recreates those monitors in your DownForReal account for you.
Export from UptimeRobot. You download your monitor list from your existing UptimeRobot account using the export option in that service.
Upload the file to us. You add the exported file during signup or from your dashboard after signup, and we show you exactly what will be created before anything changes.
Review and run both services side by side. You keep your UptimeRobot account running during your trial, so you can compare alerts from both services before you decide to cancel anything.
4 of 4 monitors imported · Alert contacts carried across for review
Pricing
There is one plan, and the plan is priced per account. You do not pay per person on your team, and you do not pay extra for a status page. A responder, a term that some competitors use in their pricing, means one individual person who receives and acts on alerts.
Billed monthly. You can cancel at any time.
The trial runs for 14 days, and it gives you the full service so that you can judge it on real alerts.
The annual price works out to $6.25 per month. You pay once each year, and the lower price is our way of thanking customers who commit for the year ahead.
Our costs grow slowly as we add customers because each customer runs a small and efficient instance of proven open source software. The cost section further down this page shows our estimated numbers in full, including the margin that we keep.
Because the economics work at $10, we have no reason to add per person charges or per feature charges later in the checkout process.
Comparison
The table below compares the published entry pricing for each service at the time this staging page was prepared. Competitor plans change over time, so you should confirm the current price on the competitor website before you make a final decision.
| Service | Monthly price | What the price covers | Charges to watch for |
|---|---|---|---|
| DownForRealThis service | $10 flat, or $75 per year | One flat price per account. The price includes the hosted monitors, the three region confirmation system, and a public status page. | There are no per responder charges and no separate status page charges in this pricing model. |
| UptimeRobot | $12 Solo plan | The Solo plan at $12 per month covers 10 monitors for a single user setup. | Costs rise as your needs grow. The Team plan costs $39 per month, and the Scale plan costs $79 per month. |
| Better Stack | $29 per responder | The starting price is $29 per month for each responder on the account. A responder is one person who receives and acts on alerts. | You pay extra charges on top of the per responder price. Extra charges apply per block of monitors and per status page. |
A worked example may help you read the table. A team with three people who respond to alerts would start at $87 per month on the per responder model described above, before any extra monitor or status page charges are added. The same team pays $10 per month here, because our price is charged per account.
Cost transparency
Most software companies treat their costs as a secret, and that secrecy makes it hard for you to judge whether a low price is sustainable. We would rather show you the arithmetic, because a monitoring service that loses money will eventually cut corners or close, and either outcome leaves you without alerts.
The figures on this page are estimates that we prepared for planning at a scale of 100 paying customers. They are labelled as estimates because real bills vary from month to month with usage, exchange rates, and provider pricing. We will update the figures as real operating data replaces the planning estimates.
Estimated infrastructure cost per month when the service supports 100 customers.
Estimated total cost for each customer each month, including payment processing.
Estimated gross margin at the $10 monthly price. Gross margin means the share of revenue that remains after the direct costs of providing the service are paid.
Giving back upstream
Upstream means the original project and the people whose work a later service builds upon. In this case, the upstream project is Uptime Kuma, and the person behind it is its author, Louis Lam.
Once revenue covers operating costs, we sponsor Uptime Kuma's author, Louis Lam, at $25 a month. The sponsorship is a flat $25 USD per month, paid through GitHub Sponsors, and it starts in the first month where subscription revenue exceeds the service's total monthly operating costs.
The amount stays at $25 a month while the service is small, and we will revisit it if monthly recurring revenue ever passes $1,000. GitHub Sponsors is the platform the author uses to receive sponsorship for the project, and sponsorship means regular financial support given through that platform.
A fork is a separate copy of a software project that develops in its own direction while it keeps the history and licence of the original. Our fork of Uptime Kuma stays public, so anyone can read the changes that we make.
The fork remains available under the MIT licence. The MIT licence is a widely used open source licence that allows people to use, copy, modify, and share the software, provided that they keep the copyright and licence notices in place.
DownForReal is an independent hosted service that is powered by Uptime Kuma. The service is built on a fork of that open source project, and the service is operated by us. Nothing on this page should be read as a claim of official status, endorsement, or partnership with the Uptime Kuma project or with Louis Lam, because no such relationship exists.
Frequently asked questions
The service has no official status. DownForReal is an independent service that is powered by Uptime Kuma, which means that it runs a fork of the Uptime Kuma open source software on infrastructure that we operate. The Uptime Kuma project and its author, Louis Lam, do not operate this service and do not endorse it.
A monitor that lives on the same server shares a failure domain with that server. When the server fails, the monitor fails at the same time and it cannot send you an alert. A hosted monitor keeps running during your outage, so it can confirm the problem, notify you, and keep your public status page available for your customers.
Confirmation means that the same address is checked from three geographic regions before a down alert is sent. If one region cannot reach your service and the other two regions can reach it, we treat the result as a local network problem and we send no alert. If all three regions cannot reach your service, we treat the outage as real and we notify you.
This rule reduces false alarms, and false alarms matter because people learn to ignore a system that cries wolf.
The trial lasts for 14 days and it gives you access to the full service. You can add your monitors, set up your alerts, publish your status page, and import your monitors from UptimeRobot during the trial period. At the end of the trial, you can move to the $10 monthly plan or the $75 annual plan if the service has earned your business.
For a typical account, the import itself takes about 2 minutes. You export your monitor list from UptimeRobot, you upload that file to us, and the importer recreates your monitors in your account. We show you a preview before the import runs, and we suggest that you keep both services running during your trial so that you can compare them directly.
You may monitor only systems that you own or systems that you are authorized to test. Authorization means that the owner of the system has given you clear permission to check it. Please do not add addresses that belong to other people or other companies without that permission, because repeated checking of a system without permission can burden that system and can create legal problems for you.
We publish the numbers because you are trusting this service with a critical job, and you deserve evidence that the price can sustain the service. Our planning estimates show a total direct cost of about $2.04 for each customer each month at the $10 price, which leaves an estimated gross margin of about 80 percent to fund development, support, and upstream sponsorship. The figures are estimates, and we describe them as estimates wherever they appear on this page.
Start your free trial
Tell us who you are and give us the first address that you want watched. That first monitor is the automated check described at the top of this page, and it gives you a real alert path that survives a failure of your own server.