How do I run changedetection.io with Docker?
Run the ghcr.io/dgtlmoon/changedetection.io image with a volume mounted at /datastore and port 5000 published, then open it in a browser. The project's docker-compose.yml does the same.
Run website change detection on your own server. Add a page, pick the part to watch, and changedetection.io alerts you when its text changes, the price drops or the item is back in stock. Alerts go out by email, Discord or a webhook, and more services are supported.
The key facts for ChangeDetection.
| Requirement | ChangeDetection |
|---|---|
| Price | Free + Paid |
| Web UI | Yes |
| Architecture | x86 + ARM |
| Containers | 1The app alone |
| Database | file only |
| Runtime | Python |
| Docker | Yes |
| Windows | Yes |
| Python | Yes |
Run the ghcr.io/dgtlmoon/changedetection.io image with a volume mounted at /datastore and port 5000 published, then open it in a browser. The project's docker-compose.yml does the same.
Yes, with a Playwright or WebDriver content fetcher, which runs a real Chrome. The default fast fetcher does not run JavaScript.
Yes. Turn on the re-stock and price detection option for a single product page, then set an upper or lower price, or a percentage change, to alert on.
More apps in Monitoring.
Better Status Page is a self-hosted status page and uptime monitor that fits on one small server. It checks HTTPS, ping, DNS, database and Docker targets and alerts you through Discord, Slack, ntfy and more.
Send alerts from scripts, apps and services through a lightweight push notification server you host.
Monitor CPU, memory and disk use across your machines from one dashboard. A small agent runs on each system and reports to a hub you host, which adds container stats, alerts and disk health.
Monitor servers, containers and homelab hardware every second, with ready-made dashboards and anomaly detection. It discovers what to watch automatically.
Watch Proxmox, Docker, Kubernetes, TrueNAS and vSphere from one dashboard, with a page for each platform. Pulse keeps the history, warns about failed backups and filling disks, and pings the router or printer that cannot run an agent.
Monitor your websites and services and get alerts through more than 90 notification services.
Track CPU, memory, disk and network use for every container on one host, from a small daemon with its own web page. The same numbers are served as Prometheus metrics, ready for Grafana dashboards and alert rules.
Point Glance at a YAML file of widgets, such as RSS feeds, Reddit posts and Docker container status, and it serves them as one dashboard. The binary is under 20 MB and the Docker image is about as small.
Updated