Telemetry
What Checkmate sends about your installation, why, and how to turn it off.
Overview
Checkmate sends one small, anonymous report per installation once a day. It tells us that an installation exists, which version it runs, how it was installed and how many monitors of each type it has. It contains nothing you typed into Checkmate: no URLs, names, emails or check results.
This page lists every field in the report, what we do with it and how to turn it off. Telemetry is on by default. Turning it off changes nothing else in Checkmate.
Why we collect it
Checkmate is self-hosted, so we have no way to see how it is used. GitHub stars and image downloads tell us someone was curious once, not whether Checkmate is still running a month later. The daily report fills that gap with the smallest set of facts we found useful:
- How many installations are running. An installation counts as active if it reported in the last 3 days. This is how we know whether a release grew or lost users.
- Which versions are in use. This tells us how quickly people upgrade and how long we need to support upgrades from older versions. If installations go quiet right after a release, something in that release may be broken.
- How Checkmate is installed. Docker image or source, single process or separate worker. We test and document the setups people use first.
- Operating system, CPU architecture and Node.js version. This shows whether ARM builds (Raspberry Pi and similar) matter and when we can raise the minimum Node.js version without stranding anyone.
- How many monitors of each type. This shows which monitor types people rely on, whether a new type gets used and how large installations get, which we use as a target for performance testing.
- Country and city. A rough sense of where Checkmate runs, which helps us decide which translations to prioritize.
What the report contains
This is a complete example of one report:
{
"instanceId": "3f0c7a52-9b1e-4d6a-8c2f-5e7d1a0b9c34",
"version": "4.0.0",
"deployment": "docker",
"queueMode": "primary",
"nodeVersion": "24.4.0",
"os": "linux",
"arch": "x64",
"monitorTypeCount": [
{ "type": "http", "count": 12 },
{ "type": "ping", "count": 3 },
{ "type": "pagespeed", "count": 0 },
{ "type": "hardware", "count": 2 },
{ "type": "docker", "count": 0 },
{ "type": "port", "count": 1 },
{ "type": "game", "count": 0 },
{ "type": "grpc", "count": 0 },
{ "type": "websocket", "count": 0 },
{ "type": "dns", "count": 1 },
{ "type": "heartbeat", "count": 4 },
{ "type": "unknown", "count": 0 }
]
}| Field | What it is |
|---|---|
instanceId | A random ID generated once and stored in your Checkmate database. It isn't derived from your hardware, hostname or network. A fresh database gets a new ID. |
version | Checkmate version |
deployment | docker for the official image, source otherwise |
queueMode | primary or worker, from the QUEUE_MODE setting |
nodeVersion | Node.js version |
os | Operating system, for example linux |
arch | CPU architecture, for example x64 or arm64 |
monitorTypeCount | The number of monitors of each type. Counts only: no names, URLs or settings. |
What we add when it arrives
The report is received by a small service we run on Cloudflare. Two things are added on our side:
- The date it arrived, so we can tell active installations from old ones.
- Country and city, which Cloudflare works out from the connection's IP address. For an installation in a cloud data center, this is the data center's city, not yours.
The IP address itself is never stored. It is used only to limit how often one address can send reports, and request logging is turned off so it doesn't end up in a log either. Location is only shown as totals (installations per country, installations per city), never next to an installation's ID.
What is never sent
- URLs, hostnames or IP addresses you monitor
- Monitor names, settings or check results
- Incidents, status pages or maintenance windows
- Users, emails, team names or passwords
- Notification channels and their settings
- Environment variables, keys or other secrets
- Logs or error messages
Who can see it
The data sits in a database run by Bluewave Labs, the team behind Checkmate. Only Checkmate maintainers can open the dashboard built on it. It shows totals: active installations, versions, install methods, platforms, monitor types and a map of countries and cities. Where it lists the largest installations by monitor count, each one appears only as the first 8 characters of its random ID.
We don't sell or share this data. If we publish numbers, for example in a release note, they are totals.
How it is sent
- Once every 24 hours, as an HTTPS
POSTtohttps://checkmate-telemetry.bluewave-labs.workers.dev/v1/heartbeat. The first report goes out at a random time within an hour of startup. - With a 5-second timeout and no retries. A failed report is skipped until the next day.
- Failures are only logged at debug level and never affect monitoring.
On startup, Checkmate logs Anonymous telemetry enabled, set TELEMETRY_ENABLED=false to disable so you know it's on.
Turning it off
Set this on every Checkmate server container (the API and any workers), then restart:
TELEMETRY_ENABLED=falseThe scheduled report is removed the next time Checkmate starts, and nothing more is sent.
Seeing what is sent
The code that builds the report is open source: telemetry.service.ts. To see the exact report from your own installation, point TELEMETRY_URL at an endpoint you control (a local request inspector works) and watch what arrives:
TELEMETRY_URL=https://your-endpoint.example/heartbeatTELEMETRY_URL is also how you would send reports to your own collector instead of ours.