Monitoring & Checks¶
A target is a monitored host with a name, optional group, active flag, interval (5-86,400 s), retention (1-3650 days), maximum retries (0-10), and individually switchable checks.
A group can only be deleted when no target uses it (HTTP 409 otherwise). Targets are
never moved to "Ungrouped" implicitly; reassign or delete them first.
Creating targets¶
A four-step wizard suggests checks by profile and shows a summary before saving.
| Profile | Suggested checks |
|---|---|
| Website | HTTP GET, success codes 200-399; TLS expiry for HTTPS |
| API / health endpoint | HTTP GET, success codes 200-299; TLS expiry for HTTPS |
| Server / host | ICMP ping |
| TCP service | TCP to the ports you enter |
| Custom | Any combination and advanced settings |
Defaults: 60 s interval, two retries, 365 days retention, auto-close. You can
change them before creation. A domain without scheme suggests HTTPS; an explicit
http:// stays HTTP. The summary does not contact the target.
In the checks step, settings appear only for active checks: the TCP port field with TCP, URL, method and success codes with HTTP, and the TLS option "Ignore TLS certificate errors" with HTTP or certificate expiry. Interval, retries, retention and auto-close are always shown. The summary lists only active checks.
Exact duplicates are blocked. Different services on the same host are allowed; partially overlapping checks show a warning. Edit opens the full editor.
Check types¶
ICMP via system ping. Result: reachable and latency.
GET or HEAD, optional basic auth, success codes (default 200-399), port
override, optional certificate ignoring. Redirects are not followed. Result:
reachable, status code, latency.
Several ports per target; status and latency per port.
Remaining days of the TLS certificate. A configured certificate host and port always win.
Without them the host comes from the HTTP URL (then the target host) and the port from
the HTTP URL, then the HTTP port override, then 443 (80 for an http URL).
Host rules¶
- Hosts must be valid hostnames or IP addresses.
- Checks block loopback, link-local, and metadata addresses (SSRF protection). Private RFC 1918 addresses are allowed.
- DNS is resolved once; the check connects to exactly that address.
Scheduling¶
Checks run aligned to the interval with a small per-target jitter. Failed checks are retried with backoff. A slow check does not delay other targets. Changes apply without a restart.
Searching targets¶
The search field on the Targets page filters while you type. It matches a part of the
name, host, HTTP URL, TCP port or group name, ignoring case and accents. Groups without
a hit disappear, groups with hits open while you search, and the badge shows hits
against the total (2 / 5). Clearing the field (the clear button or Esc) restores the
groups as they were. Without a match the page says No targets match "...".
Status and SLA¶
Each target has a status per service and an overall status (up, down,
degraded, unknown), plus uptime for 24 h and 30 days. Uptime follows the overall
status: a check cycle counts as up only if every enabled check (HTTP, ping, TCP)
is up. Admins can reset a target's measurements and send a test alert.
SLA targets (availability %, HTTP response time, ping latency) are set per target. They affect charts and reports only and never raise alerts. Without an availability target, charts assume 99.9 %. See Reports.
Dashboard¶
The dashboard shows online, offline, disabled, and total targets, the connection quality widget, and the Downtime Overview (24h). It refreshes every 30 s and has a refresh button.
On desktop, Edit dashboard moves, resizes, and hides widgets per user. Done saves the layout; Reset layout restores the default. Keyboard: arrow keys move a focused widget handle, Shift+arrow keys resize.
Live updates¶
Open pages receive change hints over a Server-Sent Events stream (/ui/events)
and reload the affected parts. If the stream fails, the page polls every 15 s. A
Server offline overlay appears only when /health fails twice in a row.
Proxy requirements: Reverse proxy.
Retention and rollups: Storage & retention.