We watch the hosts while you watch the sites.

Uptime checks, certificate expiry, and alert routing are included on every plan and managed from your own dashboard. You add the sites, you set who gets told when one stops answering, and the 99.5% uptime commitment underneath them is backed by active monitoring on each host we operate.

Add a MonitorTalk to an Engineer
carpathian.ai / dashboard / monitoring

Monitored Sites

3 monitored
carpathian.aiUp
store.carpathian.aiUp
api.carpathian.aiUp
TLS certificateValid
Expiry monitoringOn
Alert recipientsSet
Sites
3
Uptime SLA
99.5%
Certificates
Valid
Alerts
Routed
Uptime checks
Included on every plan, with no monitoring tier to buy first
Certificate expiry
SSL monitoring on the sites you add, so a renewal date stops being an outage
Alert routing
You decide who hears about a failing site, and you change that list yourself
99.5% uptime SLA
On cloud and dedicated services, backed by active monitoring on each host

What gets watched

Adding a site turns on the checks below. None of them is a paid upgrade, and there is no higher tier holding back the parts you would want.

Uptime checks

An uptime check asks your site whether it is answering and records what came back. When the answer stops arriving, the monitor changes state and the people on your notification list hear about it. That is the whole mechanism, and its value is that you find out from us rather than from a customer who gave up and emailed you. Uptime checks come with every plan, on every site you put under monitoring.

Certificate expiry

A certificate that quietly reaches its expiry date takes a site down as completely as a dead host, and it does it on a schedule anyone could have seen coming months in advance. It is the most avoidable outage there is and it still happens constantly, usually on a subdomain nobody remembered was sitting on its own certificate. SSL monitoring covers expiry on the sites you add, so the certificate becomes something you get told about instead of something you learn from a browser warning screenshot.

Alert routing

An alert nobody reads is the same as no alert. You manage the notification list from your dashboard, so the people who need to know about a given site are the ones the message reaches, and you can correct that list the day someone changes roles rather than the week after the next outage teaches you it was wrong.

Uptime checks
On every plan, on every site you add
SSL monitoring
Certificate expiry watched alongside the uptime check
DDoS mitigation
Included on every plan alongside the checks
99.5% uptime SLA
Cloud and dedicated services, backed by host monitoring

Separate controls for separate jobs

Monitoring is not one switch. Viewing the dashboard, adding a site, changing the settings on a site already there, removing one, and managing who gets notified are five separate permissions, and they can sit with five different people.

That split does more work than it sounds like it does. The person who knows which hostnames deserve a check is often not the person who should be able to switch a check off. The list of people who need to hear about a failing checkout page is rarely the same list as the people who administer it. Keeping removal apart from creation means someone can extend your coverage without also being able to quietly delete it.

Changes are recorded. Human actions, security events, and errors are logged, so who removed a monitor last Thursday is a lookup rather than a conversation.

  • Add a site to monitoring without holding the right to remove one
  • Manage monitoring settings without owning the notification list
  • Keep alert recipients current as the team changes
  • Read the dashboard without the ability to change anything on it
carpathian.ai / dashboard / monitoring access
  • View
    See the monitoring dashboard and the current status of every site under it
  • Add
    Put a new site under monitoring
  • Configure
    Change monitoring settings on the sites already there
  • Remove
    Take a site out of monitoring
  • Notify
    Manage who receives alerts and keep that list current

About our 99.5% commitment

An availability figure is the cheapest thing on a hosting page to print, so here is the reasoning behind ours instead of only the number.

Our uptime commitment on cloud and dedicated services is 99.5%, backed by active monitoring on each host. We publish a floor we intend to hold on hardware we own, in buildings we can walk into, rather than the highest number the industry would let us print.

A committed floor you can verify is worth more than a higher number you have to trust. Decimal places cost nothing to add. The questions that decide your availability are who owns the machine, who can physically reach it, and what happens in the hour after it stops responding.

Then there is the remedy. An availability guarantee is generally settled in service credits, and a credit is calculated against what you paid for the service, not against what the downtime cost your business. A larger credit on a smaller base is a different promise from keeping your site up, which is why the clause is worth reading before you compare two figures on two marketing pages.

So the number stays where we can hold it, and we would rather spend the conversation on how we hold it than on how it compares.

carpathian.ai / infra / us-central-001
US Central 001Production
US Central 002Disaster recovery
Power and coolingRedundant
Physical accessOur own staff
RacksMonitored around the clock

Who responds when a check fails

Both of our facilities are in Des Moines, Iowa. US Central 001 runs production and US Central 002 is the disaster recovery site. Both have redundant power and cooling, and the racks are monitored around the clock.

Physical access is by Carpathian's own staff. When a machine needs hands on it, the hands belong to someone who works here and has seen your cabinet, not to a contractor dispatched into a building they have never been inside. That is the half of an availability commitment no contract language can supply.

The redundancy between the two sites is site redundancy, not geographic. Both buildings are in the same metro, and we say so plainly because it changes what you should plan for.

If you want to know how coverage works outside business hours, ask us and you will get a specific answer about who is reachable and how. We would rather tell you exactly that than imply a room full of people watching screens at three in the morning.

  • Two facilities in Des Moines, Iowa, both operated by us
  • Redundant power and cooling in each
  • Physical access by Carpathian staff
  • Site redundancy between the two, never described as geographic

Questions we get asked

What is included with monitoring?
Uptime checks, SSL monitoring, and DDoS mitigation are included on every plan. There is no monitoring tier to buy and no add-on to purchase before any of it starts working.
Who can add or remove a monitored site?
Whoever you give those rights to. Adding a site and removing one are separate permissions, so someone can extend your coverage without also being able to take coverage away.
Who receives the alerts?
The people on the notification list you manage. Managing notification recipients is its own permission, separate from managing the monitors, so the person who adds a site is not automatically the person every alert goes to.
Does monitoring watch certificate expiry?
Yes. SSL monitoring covers the sites you add, so an approaching expiry date is something you get told about rather than something a customer reports as a browser warning.
What does the 99.5% uptime SLA cover?
Cloud and dedicated services, backed by active monitoring on each host we operate.
Why commit to 99.5% rather than a higher number?
Because it is a floor we intend to hold on hardware we own and staff ourselves. An availability guarantee is generally settled in service credits calculated against what you paid for the service, not against what the outage cost your business, so a higher advertised figure is worth less than it reads. We would rather publish a number you can hold us to and answer questions about how we hold it.
Who responds when something breaks?
Our own people. Both facilities are in Des Moines, Iowa, physical access is by Carpathian staff, and the racks are monitored around the clock. Ask us how coverage works outside business hours and you will get a specific answer.
Is there a record of changes to monitoring?
Yes. Human actions, security events, and errors are logged, so who added a monitor or changed the notification list is a lookup rather than a conversation.

The rest of what sits under these checks

Monitoring watches what runs on our hardware. The virtual machines themselves, the backups that make a bad day recoverable, and the buildings the racks sit in each have a page of their own.

VPS HostingBackupsData Centers

Add a monitor before you need one.

Uptime checks, certificate expiry, and alert routing come with the plan you already have. Add a site from your dashboard, put the right people on the notification list, and the first time you hear about an outage it will be from us. If you would rather talk through how the 99.5% commitment and the response behind it apply to your setup, the person on the other end runs the hardware.

Add a MonitorTalk to an Engineer