Skip to main content

Uptime

Watch websites, ports, hosts and services, get alerts when they go down, and publish status pages.

MonitoringSourceAPIChangelog
  • Check websites, TCP ports, ping, DNS records and TLS certificates
  • Watch Termix hosts, Docker containers, systemd units and your own commands over SSH
  • Run any network check from one of your hosts, to reach private networks
  • Push monitors for cron jobs and backup scripts
  • Uptime numbers, response time charts and an incident history
  • Alerts through the Alerts plugin, on every channel you set up there
  • Maintenance windows that stop false alarms
  • Public status pages with notes
  • Share monitors with users and roles
  • Dashboard cards, a panel beside the terminal and tiles on the host page

Uptime watches websites, ports, hosts and services and tells you when something goes down. It keeps a history of every outage, stays quiet during planned work and can publish a public status page.

Add a monitor​

Open Uptime from the rail and press Add monitor. Pick a type, fill in what to check, and press Test to try it before you save.

TypeWhat it checks
Website (HTTP)Calls a URL and checks the status code, and a keyword if you set one. An https URL also tracks the certificate
TCP portThat a port accepts connections
PingThat an address answers ping
DNS recordThat a record exists, and can check its value
TLS certificateThat a certificate is trusted and how long it has left
PushThat a job called its URL in time
Termix hostA host's port, the same way as the host status dot. It never logs in
Docker containerThat a container is running and not unhealthy, over SSH
systemd unitThat a unit is active, over SSH
CommandThat a command exits with 0, over SSH

To watch many hosts at once, press Add from hosts and tick them. A host that already has a host monitor is skipped.

How a monitor goes down​

A failed check doesn't alert right away. The monitor shows Retrying and checks again at the retry interval. Once the number of retries is used up it goes down, an incident opens and the alert goes out. The next good check closes the incident and sends a back up alert.

Set Remind me every to get another alert while something stays down. 0 sends one.

Check from a host​

The Termix server can only test what it can reach. Website, TCP, ping, DNS and TLS monitors can run from one of your hosts instead. Pick it under Check from, and the check runs there over SSH. This reaches services on a private network with no port open to Termix, and shows what a network sees from the inside.

The host needs the tool for the check: curl for websites, bash and timeout for TCP, ping, dig or getent for DNS, and openssl for TLS. If one is missing the monitor tells you which.

Push monitors​

A push monitor gives you a URL. Call it at the end of a cron job or backup script. If no call arrives within the interval plus the extra wait, the monitor goes down.

0 3 * * * /usr/local/bin/backup.sh && curl -fsS https://termix.example.com/plugin-api/uptime/push/YOUR_TOKEN

The URL takes optional values:

  • status=down reports a failure, so a job can say what went wrong
  • msg= sets the message
  • ping= sets a response time in milliseconds
curl -fsS "https://termix.example.com/plugin-api/uptime/push/YOUR_TOKEN?status=down&msg=disk%20full"

Copy the URL when you add the monitor, since it is only shown once. Termix keeps a hash of it. New push URL on the monitor makes a new one and stops the old one.

Alerts​

Alerts go through the Alerts plugin, so they reach every channel you set up there. Route them by category: uptime.down, uptime.up and uptime.cert_expiring.

You get alerts for monitors you add. If someone shared a monitor with you, turn alerts on with the bell on the monitor. Without the Alerts plugin, monitors still run and keep their history.

Certificates warn at the number of days you set, and again at 30, 14 and 7 days where those are lower.

Maintenance​

A maintenance window stops false alarms during planned work. Failed checks in a window show as Maintenance, open no incident and send no alert. They don't count against uptime.

A window runs one time or every week. It covers all your monitors, the ones you pick, or the ones with a tag.

Status pages​

A status page shows people how your services are doing, with no login. You need the Publish status pages permission.

  1. Open Status pages and press Add page.
  2. Give it a title and an address, which is the end of the link.
  3. Add groups and pick the monitors for each.
  4. Turn on Published and save.

A page shows each monitor's name, state, 30 day uptime and a bar of the last 90 days. It never shows addresses, error text or anything else about how a monitor works. Choose names you are happy to show. Response times only show if you turn them on.

Post a note on a page to tell people about an outage or planned work. Notes stay on the page until a week after you resolve them.

The page lives at /plugin-api/uptime/status/<address> on your Termix address. Its data is also at the same address plus /data as JSON. To put a page on its own domain, add a route to that path in the Reverse Proxy plugin.

An admin can turn off every status page at once with Allow public status pages.

Sharing​

Monitors belong to the person who added them. Open a monitor and press Share to give a user or role view or edit access. People you share with see the live state, the history and the settings, but never saved logins or headers. People with view access can't see a stored command.

Only the owner can share or delete a monitor.

Checks run as the person who last changed what the monitor looks at. If they lose access to a host, the monitor goes down and says so.

Private addresses​

Only people with the Check private addresses permission can point a monitor at a LAN, Tailscale, loopback or other private address from the Termix server. It is on for admins by default. This is so a normal user can't use Termix to look around the network it sits on.

The check happens on save and again on each run, so a name that later starts pointing at a private address gets refused too.

Check from a host doesn't need this permission, since you are using a host you already have access to.

Permissions​

PermissionDefaultWhat it allows
View monitorsadmin, userSee monitors, history and incidents
Manage monitorsadmin, userAdd and change monitors and maintenance windows
Check private addressesadminPoint monitors at private addresses
Run commandsadminAdd monitors that run a command on a host
Publish status pagesadminMake status pages anyone with the link can see

History​

The last 1440 checks of each monitor are kept for the bars and the 24 hour chart. Older checks are added up per hour for a week and per day for as long as Keep history for says. Keep history for also limits how long incidents stay.

Single checks don't go in the database. They live in memory and in a small file in the plugin's data folder that holds only numbers, so frequent checks don't slow down Termix.

Ping in Docker​

Ping needs the ping program on the Termix server. The image has it from Termix 26.11.1. On anything older, and on servers that aren't Linux, use a TCP monitor or Check from a host.

Automations, Fleets and the assistant​

Uptime adds actions and events to Termix's Operations layer:

  • Automations can start when a monitor goes down, comes back up or a certificate is about to end
  • Automations, Fleets, the CLI and the phone app can check, pause and resume monitors
  • The assistant can read the summary and the monitors on a host
  • The host page shows a tile with the monitors for that host