Uptime monitoring

Know it's downbefore anyone tweets it.

Checkmate checks your websites and APIs from 6 continents, validates what they return, then pages the person who owns the monitor. All of it runs on your own servers.

checkmate.soUp
https://checkmate.so · checked every 60 s
Uptime / 30 days
99.98%
2 incidents
Avg response
142 ms
p95 380 ms
Cert expiry
84 days
Mar 12, 2027
Domain expiry
213 days
Jul 19, 2027
Response times
DayWeekMonth
00:0006:0012:0018:00Now
Recent checksFrankfurt · London · Virginia
12:04:32200 OK138 msFrankfurt
12:03:31200 OK129 msVirginia
12:02:32200 OK151 msLondon
11:56:12504 Gateway timeout30,000 msFrankfurt
Global checks

Checked from 6 continents, not one datacenter.

A site that loads fine from Frankfurt can be unreachable from Sydney. Checkmate runs distributed checks through the GlobalPing network, so a regional outage shows up as a regional outage instead of a mystery.

Every check location is plotted on an interactive map with its latest result, which makes 'is it just me?' a question you can answer in one glance.

  • Checks from locations on 6 continents
  • Interactive map with per-location results
  • Works alongside checks from your own instance
Distributed uptimeGlobalPing
Frankfurt38 ms
N. Virginia92 ms
São Paulo148 ms
Singapore171 ms
Sydneymalformed response
Cape Town204 ms
6 locations · checked every 60 s · slowest region highlighted
Response validation

A 200 response can still be wrong.

Your API can return an empty body, a stale cache or an error page with a cheerful status code. Checkmate goes past the status line: match keywords in the response, or query JSON bodies with JMESPath or a regex and fail the check when the answer is wrong.

That turns uptime monitoring into a basic correctness check for every endpoint your customers depend on.

  • Status code checks
  • Keyword and content matching
  • JSON validation with JMESPath or regex
Response validationJSON path
Match type
JSON path (JMESPath)
Expression
services.payments.status
Expected value
ok
Last check passed · value was "ok" · 200 in 118 ms
Alerting

Alerts that wait for proof.

One dropped packet should not wake anyone up. A monitor can require that the last n of m checks fail before it flips to down, so alerts fire on real outages instead of blips.

When a monitor does go down, Checkmate opens an incident with a timeline you can acknowledge and reviews later. Notifications go out on any of 12 channels, from email to PagerDuty.

  • n of m confirmation before a status change
  • Incidents with timelines and acknowledgements
  • 12 notification channels
Incidents1 active
Last 30 days
Open now
1
api.checkmate.so
Resolved / 30d
4
across 3 monitors
Avg duration
26 min
median 9 min
Longest
1 h 12 m
storage failover
Incident days4 of the last 30 days
30 days agoToday
Timeline
Activeapi.checkmate.so · HTTP 50412 min
HTTP/1.1 504 Gateway Timeout · verified from Frankfurt and Virginia
12:04 detected12:07 acknowledgednow
Resolveddb-primary · connection refusedAug 12 · 22 min
Resolvedcheckout.acme.com · timeoutAug 2 · 4 min
Resolvedcdn-edge · SSL handshake failedJul 30 · 9 min
History

Patterns show up before outages do.

Response time charts span a day, a week or a month, so a dependency that got 40 percent slower on Tuesday is visible on Wednesday, not after the postmortem.

Uptime percentages are tracked to two decimals per monitor. SSL certificate expiry sits on the same page, so one view covers most of what you ask about an endpoint.

  • Day, week and month chart views
  • Uptime percentage to two decimals
  • Certificate expiry on the same monitor
api.example.com
DayWeekMonth
Uptime / 7 days
99.94%
1 incident
Avg response
131 ms
p95 402 ms
Checks
10,080
every 60 s
MonWedFriSun
Who runs this

From a single homelab box to a client fleet.

Self-hosters

Watch the services on your own hardware without shipping availability data to a third party. One compose file and the dashboard is yours.

SaaS and ecommerce teams

Downtime costs revenue and trust. Validate the endpoints your checkout and API customers hit. Failures reach you before support tickets do.

Agencies and MSPs

Monitor every client site from one dashboard with no per-monitor pricing. Tags keep a large fleet sorted by client or project.

Scope

What it covers, and what it doesn't.

Covered

  • HTTP and HTTPS endpoints, with status code checks
  • Keyword and content matching on responses
  • JSON response validation with JMESPath or regex
  • Response time history with day, week and month views
  • SSL certificate expiry on the same monitor
  • Distributed checks from 6 continents via GlobalPing

Out of scope

  • Real-user monitoring: these are synthetic checks, not analytics from visitor browsers
  • Application tracing and profiling: pair Checkmate with an APM tool for code-level timings
  • Log aggregation for your applications
Under the hood

For the technically curious.

Intervals you control

Set the polling schedule per monitor, from short intervals to custom schedules. Smart throttling backs off when a monitor has been stable, so a large fleet stays cheap to run.

Status changes need evidence

Configure a monitor to flip only when the last n of m checks agree. Flappy networks stop producing flappy alerts, in both directions: down needs proof and so does recovered.

Outbound proxies

Route checks through a proxy, globally or per monitor. Useful when your instance sits behind strict egress rules or when a target only admits traffic from known addresses.

Your data stays put

Checks run from your instance and results land in your database. There is no phone-home, so the list of what you monitor is nobody's business but yours.

FAQ

Frequently askedquestions.

Uptime monitoring sends regular requests to a website or API and alerts you when it stops responding correctly. Checkmate runs these checks from your own server, records the results and notifies you the moment a monitor goes down.

Each check sends a request, then inspects the status code, optionally the response content and how long it took. You can require several failed checks in a row before the monitor flips to down, so a single network blip doesn't page anyone.

Yes. Beyond status codes, a monitor can match keywords in the response or query JSON bodies with JMESPath or a regex. If the shape or content of the answer is wrong, the check fails even when the status code is 200.

From your own Checkmate instance by default, with optional locations on 6 continents through the GlobalPing network. The distributed results are plotted on a map so regional problems are easy to spot.

Two tools. n of m confirmation flips a monitor only when the last n of m checks agree. Maintenance windows suppress alerts on a schedule while you deploy or patch.

Not for uptime monitoring. Checks run from the Checkmate server itself. The Capture agent is only needed if you also want hardware metrics like CPU, memory and disk from your machines.

Yes. Checkmate is open source under AGPL-3.0. Self-host it and every feature is included, with no monitor limits or per-check pricing.

Get started

Every feature,no paywall.

Checkmate is open source under AGPL-3.0. Self-host it and this feature ships free, on your servers, with your data.