Monitoring that describes the machine, not just the app
CPU, memory, disk, network, and per-container usage, sampled every fifteen seconds and streamed live — including steal time, which is how you find out a slow application is a noisy neighbour rather than your code.
What you get
Steal time
On the cheap instances most of us run, a slow app is often someone else’s workload. Nothing else on a typical dashboard tells you that.
Every container
Running, stopped, and unhealthy counts; per-container CPU, memory against its limit, network, restarts, health, and out-of-memory kills.
Live, then history
Samples stream live while you watch. One row a minute is kept, rolled up hourly after seven days, for ninety days in all.
Disk that cleans itself
A nightly sweep reclaims build cache, old images, and stale source trees — and never touches a volume, which is your data.
In detail
- Host facts: operating system, kernel, architecture, uptime, public address, and virtualisation type.
- CPU: model, cores, total and per-core usage, load averages, and steal.
- Memory and swap; per-filesystem usage and inodes; Docker’s own disk use broken out into images, containers, volumes, and build cache.
- Per-interface network throughput and rates.
- Live application logs from the running containers.
- Run the storage sweep on demand and see what each of the last ten runs reclaimed.
Works well with
Alerts
Alerts that open and resolve themselves — in the portal, email, webhook, or Telegram.
Request analytics
Every request to every app — read from the proxy, with no script to install.
Your own servers
One command connects a VPS you already own. Nothing inbound is opened.
Your server is ready when you are
Create an account, connect a VPS, and ship your first application.