Tool in use

Uptime Kuma

monitoring for websites, APIs and servers. Free, open source, GDPR-compliant. What it is, how we use it, and what a tax firm watches with it day and night.

Project profile

Uptime Kuma

A fancy self-hosted monitoring tool

As of: May 30, 2026

GitHub stars

87k

Forks

7.9k

Open issues

752

License

MIT

Latest version

v2.3.2

Language

JavaScript

First release

July 3, 2021

Last commit

May 30, 2026

→ GitHub repository → Official website → Documentation → Live demo

Third-party source · Wikidata (CC0)

Wikidata profile

Uptime Kuma

Q118611125

License

MIT License

→ Wikidata entry

What Uptime Kuma is

is an open-source monitoring tool that runs on your own server in a container. At configurable intervals (typically 20 to 60 seconds) it checks whether websites, APIs, mail servers, databases and containers are reachable and how they respond — and raises an alert when something is wrong.

Unlike cloud services like Pingdom, UptimeRobot or Better Stack, all measurement and configuration data stays on your own server. For law firms, practices, tax firms and any organisation with confidentiality or GDPR requirements, that is a meaningful difference — the monitored URLs, hostnames and response times never leave your infrastructure.

Why monitoring belongs in every server architecture

A server that is not monitored does not show up when it fails — it shows up when a customer calls. That is regularly hours too late: failed mail acceptance means lost client communication, a failed website means missed inquiries, a failed backup means in the worst case unrecoverable data loss.

Monitoring is the cheapest insurance against symptoms that get expensive — a 30-second detection is usually the difference between „a brief blip" and „half a day lost". Kuma costs nothing, runs on hardware that is already there, and is set up in half a day. There is no sensible reason to run a productive setup without such a tool.

Real-world example

How a tax firm uses Uptime Kuma

Tax firm with twelve workstations, own server in the office, client portal via , mail server via Postfix, daily backups to a NAS. These were the six requirements before introduction:

Client portal reachable at all times

Clients upload documents outside office hours — if the portal fails at night, the firm only notices on Monday morning via the complaint. That should no longer happen.

DATEV login and external services in view

When DATEV document transfer or the state tax portal stalls, every minute of waiting costs hours in the monthly close. Early warning instead of guessing on the phone.

Mail server monitored without giving it away

Our own Postfix takes confidential client communication. Cloud monitoring would write credentials and hostnames into third-party logs — that was not an option.

SSL certificates must never expire

Auto-renew via Certbot has run for years, but an expired certificate in client communication would be a GDPR-reportable breach of trust. Double safeguard required.

Backup jobs must report green every day

A non-running backup only shows up at the restore attempt. By then weeks may have passed — and backdating is not possible in the GoBD world.

Notifications GDPR-compliant, no US service

Alerts should go into the team inbox and to mobile — but not via Slack, WhatsApp Business or PagerDuty. Own mail server and a Matrix bridge were the requirement.

What Kuma now watches around the clock

Eight monitors cover the full stack — from the public client website to the backup job that runs at three in the morning. Check intervals between 60 seconds (critical) and 5 minutes (non-critical), alerts via email and push notification on the work phones:

HTTPS endpoint client portal

Fetch the login page every 60 seconds, status code 200 plus keyword check on the expected HTML element. Detects not only down status but also empty pages and broken logins.

HTTPS check DATEV document transfer

External service pinged every minute. On failure clients are informed proactively — instead of being surprised by call-backs.

IMAP and SMTP port on the mail server

Plain TCP port checks without login. See port blocks (provider change, firewall drift) before clients report bounces.

SSL certificate expiry for 6 domains

Time-to-expire monitor: alert 30 days, 14 days and 3 days before expiry — three escalation steps so someone reacts even during holidays.

Backup webhook (push monitor)

After a successful run the backup script pings a Kuma push endpoint. If the ping fails to arrive within 25 hours, Kuma raises an alert — whether crashed, storage is full or the backup tool is broken.

Docker container health status

Docker-native integration: Kuma queries the socket directly for health status per container. Crash-looping containers show up immediately, before users notice.

DNS resolution of your own domains

DNS monitor checks whether the firm-wide domain still resolves to the right IP. Provider drift, DNS hijack or accidentally wrong records become visible.

Disk usage via JSON query

Small bash script writes disk usage as JSON to a protected URL daily. Kuma reads the value and alerts from 85% fill level — before „No space left on device" appears somewhere.

What Uptime Kuma can do besides

The feature set goes well beyond pure reachability checks. Six building blocks that turn the tool from a pure ping device into a small monitoring platform:

15 monitor types out of the box

HTTP/HTTPS, TCP port, ping (ICMP), DNS, keyword match in HTML, JSON query, gRPC, MQTT, container, push (heartbeat), Steam server, SQL servers (Mongo, MySQL, , MS-SQL, MariaDB), RabbitMQ. Enough for 95 percent of all SMB setups without a plugin.

Status page for customers

Own status page on your own domain, filtered to the services customers actually use. Show client portal status publicly — internal monitoring stays invisible.

90+ notification providers

Email (SMTP), , Matrix, Telegram, Discord, Signal, Pushover, Gotify, ntfy, Apprise. If you need to stay GDPR-compliant: SMTP, Matrix , ntfy , into your own systems — all without cloud lock-in.

Maintenance windows

Planned maintenance is entered in advance — during this time Kuma suppresses alerts and shows „maintenance in progress" on the status page. No 2 a.m. alerts during a planned update.

Tags and groups

Monitors can be grouped by customer, environment or criticality. One status page per client, one overview per service tier, one escalation rule per tag.

API + backup as SQLite

REST for automation and reports. Persistent data in a single SQLite file — backup is a file copy, migration between servers is an rsync and container restart.

What would the alternative be

Paid SaaS alternatives compared

If you do not want your own setup or do not have suitable infrastructure, there are commercial SaaS vendors. Three common alternatives side by side — as of 2026:

SaaS classic

Pingdom

SolarWinds, US

· From 15 USD/mo (10 monitors)
· Enterprise plans 250 USD+/mo
· US hosting, GDPR friction
· Page speed and synthetic transactions

Entry level

UptimeRobot

Malta / US infrastructure

· Free: 50 monitors, 5-min interval
· Solo 7 USD/mo (60-sec interval)
· Team plan from 30 USD/mo
· Very popular in the SMB segment

Premium / EU-capable

Better Stack

Czech Republic / EU option

· Free: 10 monitors, 3-min interval
· Team plan from 34 USD/mo
· EU servers possible, modern UI
· Incident management integrated

For a tax firm with one or two domains and a handful of internal services the SaaS approach quickly comes to EUR 200 to 600 per year — and with US vendors GDPR friction via on top. Kuma costs a one-off setup effort and runs on hardware that is already there. In architectures with your own server or NAS, Kuma is the economically and legally most pragmatic choice. SaaS tools are worth it especially if you explicitly run no own infrastructure or need multi-region probing that a tool cannot deliver.

What it costs

Free. Open source. No backdoor.

License

MIT license. Unlimited monitors, status pages, notifications. No Pro version, no paywalled features.

Running costs

Zero euros for the tool. Resource needs: about 100 MB RAM and under 1 percent CPU on typical hardware — runs comfortably on a Raspberry Pi, mini PC or the backup server you already have.

One-off setup effort

With our accompaniment typically half a day — including container setup, all relevant monitors, notifications, a status page and a short handover to you or your IT.

The only thing you invest is one-off effort for design and setup. After that the tool runs without license costs, without cloud dependency and without data outflow — year after year.

Uptime Kuma dashboard: monitor list on the left, detail pane with uptime, response time, SSL cert expiry and response-time chart on the right. — Uptime Kuma dashboard in dark mode. Source: louislam/uptime-kuma (MIT license)

Uptime Kuma Telegram notification with service-down alert (getaddrinfo ENOTFOUND) and recovery message (Up 200 OK). — Alert via Telegram. 90+ other providers available (email, webhook, Matrix, ntfy, Discord, Signal). Source: louislam/uptime-kuma

Goes with this

If you want to hold your infrastructure yourself anyway

is one of several tools we recommend and set up for clients with sovereignty and GDPR requirements. If the bigger picture interests you — server, containers, models, knowledge base on your own hardware:

→ Your own AI server & RAG → Process automation with n8n and Camunda → All tools overview

Ready for the next step?

Free intro call, no strings attached. In 30 minutes you'll know whether and how AI can help your business.

Book a call Check eligibility