title: "Core concepts" description: "The objects you'll see in the BoxWatch dashboard — servers, metrics, checks, alerts, and the rest." last_updated: "2026-05-24"
Core concepts
This page is a tour of every object that appears in the BoxWatch dashboard. Skim it once after you install your first server — most of these become familiar within a day.
Servers
A server is a physical or virtual machine running the BoxWatch agent. Each server has a name, a hostname, an IP, and an agent key. Servers are the unit of pricing — your plan caps how many you can add.
A server is "online" when its agent has reported within the last 10 minutes. Past that, it's offline and you'll get an alert (unless a maintenance window is open).
The Server Monitoring docs are coming soon — for now, see the Quickstart to install the agent.
Metrics
Metrics are the time-series data points the agent collects: CPU usage, memory, disk per mount, 1/5/15-minute load, network in/out, and uptime. The agent always collects everything; you toggle which charts you want to see per-server.
Push interval (how often the agent reports) and retention (how long BoxWatch keeps the data) both scale with your plan. All plans stores 24 hours at a 60-minute interval. Scale only stores 90 days at a 1-minute interval.
Cron checks
A cron check is a heartbeat URL that watches a scheduled job. Your job pings the URL on success (or start, or fail). If the job doesn't ping within the expected interval plus a grace period, you get an alert.
Use this for nightly backups, hourly cleanups, weekly reports — anything that's supposed to run on a schedule. See Cron heartbeat monitoring.
Watched processes
A watched process is a process name BoxWatch will alert on if it disappears or restarts. The agent reads ps on every tick and compares against the list you configured. You'll get an alert when a watched process count goes to zero, or when its oldest PID changes (restart detection).
You can also alert on aggregate CPU or memory thresholds per process. Process monitoring docs are coming soon.
Uptime checks Pro+
An uptime check is an HTTP, TCP, or TLS probe. Unlike most uptime monitoring services, BoxWatch runs the probes from the agents you already have installed — so you can verify that Redis is reachable from web-01, not just that it answers from the public internet.
TLS certificate expiry monitoring is part of the same system. Synthetic uptime check docs are coming soon.
Server groups
A server group is a logical label like production, staging, or eu-west. Groups filter your dashboards, route alerts, and apply maintenance windows to many servers at once. A server can belong to one group at a time.
Server group docs are coming soon.
Maintenance windows
A maintenance window suppresses alerts for a server or group during a planned timeframe. Metrics keep flowing, but alert emails, Slack messages, and webhooks are silenced until the window closes.
Useful for OS upgrades, scheduled deploys, and one-off migrations. Maintenance window docs are coming soon.
Notifications
A notification channel is where alerts go. Email is always on. You can also configure Slack, Discord, and generic webhooks per-account, and override channels per-server if a specific machine should ping a different room.
Alerts & Notifications docs are coming soon.
TV dashboards Pro+
A TV dashboard is a full-screen, auto-refreshing view designed for an ops wall or a spare monitor. Each dashboard has a unique URL with an embedded key — no login required, so a Raspberry Pi or Chromecast can load it directly.
TV dashboard docs are coming soon.
Status pages
A status page is a public page (e.g. status.yourcompany.com) that shows current uptime for a selection of your servers, plus any incidents you've published. Each page has its own slug, theme, and incident timeline.
Status page docs are coming soon.
Custom API endpoints
A custom API endpoint lets you expose a slice of your metric data — specific servers, specific metrics, last hour only — at a dedicated URL with its own API key. Useful for embedding a number in a Notion page, building a Grafana panel, or feeding metrics into a script.
Custom API endpoint docs are coming soon.
Plans
BoxWatch has four plans: Hobby, Pro, Team, and Scale. Each plan sets caps on server count, push interval, retention, and the feature-specific limits called out above. See Plan tiers & limits for the full comparison.