Uptimy

Uptimy Agent vs Uptime Kuma

A lighter, cluster-native alternative to Uptime Kuma

Uptime Kuma is a great self-hosted monitor, and for a dashboard on one server it's hard to beat. Uptimy Agent is built for running next to your services instead: monitors from Kubernetes labels or YAML, CronJobs without pings, and a fraction of the memory. Here's how they compare, including where Kuma is the better choice.
Get Uptimy AgentWhat Uptimy Agent does

Measured side by side

Same HTTP monitors every 60 seconds against one local nginx, both on SQLite, on the same machine. Method, scripts, raw samples and limits.

Image download

Agent
9 MB
Kuma
603 MB
Kuma slim
180 MB

Memory, no monitors

Agent
8 MiB
Kuma
126 MiB
Kuma slim
127 MiB

Memory, 350 monitors

Agent
25 MiB
Kuma
154 MiB
Kuma slim
157 MiB

CPU average, 350 monitors

Agent
0.9%
Kuma
5.6%
Kuma slim
3.5%

Railway cost a month, 350 monitors

Agent
$0.43
Kuma
$2.63
Kuma slim
$2.24

Per monitor the two are close; most of the gap is the runtime. It matters where you pay for memory, like Railway, or run one monitor per cluster or project. Both cost little: Kuma's $2.63 a month fits a Hobby plan's included usage too.

What each does

Built for

Uptimy Agent · Running next to your services: in a cluster, a Railway project or a private network

Uptime Kuma · A self-hosted monitoring dashboard

Runtime

Uptimy Agent · One Go binary, distroless image

Uptime Kuma · Node.js

Monitor types

Uptimy Agent · HTTP(S), keyword, TCP, ping, DNS, TLS expiry, Postgres, MySQL, Redis, Kubernetes workloads and Services, heartbeats

Uptime Kuma · About 30, including all of those except Kubernetes, plus JSON queries, Docker, gRPC, MongoDB, MQTT, SQL Server, SNMP, real-browser checks and game servers

Kubernetes

Uptimy Agent · Label a Service, Ingress, Deployment or CronJob and it's monitored; read-only RBAC

Uptime Kuma · —

Cron jobs

Uptimy Agent · Heartbeats on an interval or a cron schedule with time zones, exit codes and durations. Kubernetes CronJobs need no ping

Uptime Kuma · Push monitors on an interval

Notifications

Uptimy Agent · Email, Slack, Teams, Discord, Telegram, ntfy, PagerDuty, webhooks, Uptimy, and 30+ through Shoutrrr

Uptime Kuma · About 100 built in

Status pages

Uptimy Agent · One, on its own domain if you like, with sections, logos, incident updates, notices and badges

Uptime Kuma · Several, each on its own domain, with incident posts and badges

Maintenance windows

Uptimy Agent · One-off

Uptime Kuma · One-off and recurring

Users

Uptimy Agent · Several, with admin and viewer roles

Uptime Kuma · One admin account

Monitors as code

Uptimy Agent · YAML (file, ConfigMap or env var), or Kubernetes labels

Uptime Kuma · In the UI (third-party tools add config files)

API

Uptimy Agent · REST API with tokens

Uptime Kuma · Push URLs, badges and Prometheus metrics; managed through the UI

Two-factor sign-in

Uptimy Agent · Yes, with recovery codes; admins can reset it for others

Uptime Kuma · Yes

Prometheus metrics

Uptimy Agent · Yes, plus a ServiceMonitor in the Helm chart

Uptime Kuma · Yes

Storage

Uptimy Agent · SQLite

Uptime Kuma · SQLite or MariaDB

When the monitor itself is down

Uptimy Agent · Optional: Uptimy alerts you from outside when the agent stops checking in

Uptime Kuma · Needs a second monitor

License

Uptimy Agent · Apache-2.0

Uptime Kuma · MIT

Compared against Uptime Kuma 2.5.5, October 2026.

Which to pick

Uptime Kuma is the better fit if

  • You need monitor types the agent doesn't have: Docker containers, gRPC, MongoDB, MQTT, game servers, real-browser checks.
  • You want several status pages, say one per product or customer.
  • You rely on a notification service that Uptimy doesn't cover.

Uptimy Agent is the better fit if

  • You run Kubernetes and want monitors to come from labels in your manifests.
  • You want CronJobs and cron jobs tracked with schedules, exit codes and durations.
  • You want monitors in Git (YAML) and a REST API, or several users with roles.
  • You run one monitor per cluster, project or network and want it small: 9 MB, 8 MiB idle.

Switching takes a few minutes

The agent imports Kuma's database: monitors with their intervals, retries, headers and auth, notification channels and which monitors alert where, and what was on your status page.
Terminal
# 1. Copy Kuma's database (stop it first, so the copy is complete)
docker stop uptime-kuma && docker cp uptime-kuma:/app/data/kuma.db .

# 2. Start the agent
docker run -d --name uptimy-agent -p 8080:8080 -v uptimy-agent:/data ghcr.io/uptimy/agent

# 3. Sign in (password in: docker logs uptimy-agent), then
#    Settings → Import from Uptime Kuma → upload kuma.db → review → import

Frequently Asked Questions

Yes. Under Settings → Import from Uptime Kuma, upload kuma.db (stop Kuma first so the copy is complete). You see what each monitor and notification becomes and choose what to import. Check history isn't imported, and Kuma installs on MariaDB aren't supported yet.

Types the agent doesn't have, such as Docker, gRPC and inverted checks. They're listed with the reason, so nothing disappears silently. Push monitors become heartbeats with new ping URLs, so update the jobs that call them.

Uptimy Agent 0.1.7 and Uptime Kuma 2.5.5 (default and slim images, both on SQLite) ran side by side in Docker with the same HTTP monitors every 60 seconds against one local nginx, with the load checked in its access log. Memory and CPU are docker stats samples every 5 seconds over 5 minutes. The scripts, raw samples and limits are in the repository.

That's not the claim. Per monitor the two cost about the same; going from 50 to 350 monitors added 13 MiB to the agent and 22–23 MiB to Kuma. The difference is the baseline: a Go binary against a Node.js runtime. It matters most where you pay for memory, or run one per cluster or project.

If Kuma does what you need on one server, there's no pressing reason. Switch if you want monitors in Git or from Kubernetes labels, CronJob tracking without pings, several users, or something light enough to run in every cluster or project.

Looking at hosted monitoring instead? Uptimy vs Uptime Kuma covers incidents, workflows and checks from many regions.

23:00 · goodnight

All systems operational

Try it next to your services.

Uptimy Agent is open source and runs in a minute, with your Kuma monitors imported.

Get Uptimy AgentKubernetes guide