Sheltie keeps watch over your Docker homelab: websites, containers, network failover and host health. It tells you in plain words what is wrong, fixes what it safely can, and shows what happened while you were away.
Free under Apache-2.0 · No telemetry · Runs on a Raspberry Pi
Sheltie records when the internet went down and when it came back. Open the dashboard later and “Since your last visit” tells you the whole story.
Planned: alerts queue up during the outage and arrive as one digest when you are back online.
When the host comes back, Sheltie announces it booted, restarts containers that crashed, and leaves the ones you stopped on purpose alone.
Planned: on boot it will say whether it was a power cut, a crash or a reboot, and for how long.
Pair Sheltie with an external push monitor (for example Uptime Kuma or healthchecks.io) to be told when the whole host stops answering.
Planned: a built-in heartbeat and an off-site Sheltie Sentinel with the last known state of your lab.
Everything a homelab needs to stay watched, in one small container.
Every page opens with a plain sentence: “2 problems need attention”, “7 of 8 sites up”. The numbers come after.
Open Sheltie after a day away and see exactly what happened: outages, recoveries, crashed containers and automatic repairs.
Problems come with the button that fixes them: start a crashed container, open the failing site, silence alerts for an hour.
Get told when containers die. Sheltie restarts the ones that crashed, and leaves alone the ones you stopped on purpose.
Ethernet and Wi-Fi state, default-route changes and internet reachability, shown as a simple connection path.
HTTP status, latency, redirects and keyword checks with heartbeat bars, uptime, average and p95 response time.
CPU, memory, disk and CPU temperature with thresholds, durations and cooldowns so you only hear about real problems.
Quiet, state-change-only alerts, plus /status, /docker, /restart and /silence from your phone. Stale commands are ignored safely.
Bottom tabs, bottom sheets and 44px touch targets. Install it to your home screen and check your lab from anywhere.
One container with host networking and read-only Docker socket access. Add Telegram later with two environment variables.
mkdir -p sheltie/config sheltie/state && cd sheltie
curl -fsSL https://github.com/xrg360/sheltie/raw/master/config/config.yml -o config/config.yml
docker run -d --name sheltie --restart unless-stopped \
--privileged --network host \
-v /var/run/docker.sock:/var/run/docker.sock:ro \
-v $(pwd)/config:/app/config -v $(pwd)/state:/app/state \
ghcr.io/xrg360/sheltie:latestYou run a Docker homelab and want one calm place that tells you what broke on the box itself, and helps you fix it.
Read the comparisonYou want to know when something breaks in your homelab and what to do about it, from your phone.
Read the comparisonYou need to know quickly that something broke and fix it, without reading charts.
Read the comparisonYes. Sheltie is open source under the Apache-2.0 license, with no paid tier and no telemetry.
Yes. Images are published for amd64 and arm64, and Sheltie is designed to be light enough for a Pi or mini-PC.
Uptime Kuma focuses on endpoint checks and status pages. Sheltie watches the homelab host itself (Docker, network, health), repairs what it safely can, and tells you what happened while you were away. Many people run both.
No. The dashboard works on its own with in-app and desktop notifications. Telegram adds alerts and commands on your phone.
No. Sheltie tracks docker stop and compose stop and leaves those containers alone until you start them again.
No. It only talks to your own Sheltie API. No fonts, CDNs, analytics or trackers.
Sheltie is built in the open. Star the repo to follow along, or pick a good first issue and add the monitor you wish existed.