Alerts and notifications

Alerts go out through channels, added on the notify page. Each channel picks which monitors and which events it gets, and can wait or repeat. send test checks a channel before you save it.

Every alert says which monitor, what happened and, for down alerts, the cause per location and a link to the diagnostics. Emails use the same layout as the web UI, in light and dark.

Desktop notifications: on the notify page each person can also turn on browser notifications. While a pingtxt tab is open, even in the background, that browser shows when a monitor goes down or comes back up. Nothing is sent to anyone else.

Each channel has a scope and events:

• scope: all monitors (also the ones added later), or the monitors of some groups and tags, or none (only the monitors linked one by one).
• events: down (with its reminders), up (recovery, only to channels that got the alert), degraded (some locations fail or are slow, and when that ends; off by default), cert_expiry, domain_expiry and agent_offline (agent offline alerts: about agents, so it ignores the scope and the monitor links).

A monitor can also be linked to a channel one by one, from the channel (its monitor list) or from the monitor form (notifications). The link has events of its own and wins over the scope, so it adds a monitor, changes what it sends, or leaves the monitor out of an all monitors channel.

Each channel has two optional settings, in minutes:

• send after: the channel gets the alert only if the monitor is still down after this time (0 = right away). Add a second channel (the on-call phone, a manager's email) with 15 or 30 minutes to escalate long outages: short blips never reach it, and it gets the recovery message only if it got the alert.
• repeat every: the channel gets a STILL DOWN reminder at this interval while the monitor stays down (0 = once, at least 5). Webhooks carry reminder (1, 2, ...) and down_since (unix ms).

Times count from when the incident opened. What each channel received is stored with the incident, so a restart does not repeat or skip alerts.

On the maint page plan windows when alerts should stay quiet: one-off or recurring (daily, weekly, monthly, in any time zone), for some or all monitors, or whole groups and tags.

During a window the monitors show maintenance: checks keep running, but no incidents open, no alerts are sent and those checks do not count toward uptime. A monitor still down when the window ends gets its alert then.

pingtxt watches the TLS certificate of https monitors and the registration of their domain (RDAP, with WHOIS for registries without RDAP, such as .it). Turn each one on per monitor. Thresholds in days are set in settings → monitors (default 14/7/3/1 for certificates, 30/14/7/1 for domains): one alert per threshold crossed, one when it expires and one when it is renewed. Channels get them with the cert_expiry and domain_expiry events.

Periodic reports, on the notify page, email the uptime, incidents and downtime of the last week (every Monday) or month (on the 1st), per monitor group, compared with the period before. Pick the email channels, optionally some groups, the hour and the time zone. preview shows a report and send now sends one before you turn them on.

Webhooks get a JSON body with monitor_id, monitor, target, status, cause and at (unix ms), plus:

Use send test to see a real payload at your endpoint.