Uptime monitors

Uptime monitors check your websites, APIs and servers from Plenix on a schedule and tell you when one stops answering. Open them from Monitoring → Uptime Monitors in the sidebar. They come with the Monitoring module.

Kinds of check

• HTTP — requests a web address (GET or HEAD) and checks the response code. By default any 2xx or 3xx code counts as up; you can list your own codes or ranges, for example 200,204 or 200-299.
• Keyword — an HTTP check that also looks for a word or phrase on the page. Choose whether the word must be there (for example "Welcome") or must not be there (for example "Error").
• TCP — checks that a port on a server accepts a connection, for example a mail server on port 25 or a database port.

Monitors can only check public addresses. Private and internal network addresses are refused.

Creating a monitor

Click New monitor, then set:

• Name, type and the address (or host and port for TCP).
• Check every — 1, 2, 5, 10, 15 or 30 minutes, or every hour.
• Timeout — how long to wait for an answer.
• Failures before down — how many failed checks in a row before the monitor is called down. One failure is tried again a minute later, so a single dropped request is never an outage.
• Slow above (optional) — a response time above which the check counts as slow.
• Client (optional) — the CRM company the monitor belongs to, and tags.
• Alerts and Open a ticket — see below.

The first check runs within a minute.

Alerts and tickets

With Alerts on, the people set up for Uptime Monitors under Settings → Notifications are told in the app and by email when a monitor goes down and when it comes back. With Open a ticket on, a ticket is opened at the priority you choose when the monitor goes down, linked to the monitor's client. Outages that start during planned maintenance on a linked status page are marked Maintenance and do not alert.

Reading a monitor

• Uptime for the last 24 hours, 7, 30 and 90 days — the share of checks that passed.
• Avg response for the last 24 hours and 30 days.
• Certificate — for HTTPS addresses, days until the certificate expires. It turns amber at 14 days or less.
• Last 90 days — one bar per day; hover for the figure and how long it was down.
• Response time — a chart for the last 24 hours.
• Outages — every outage with start, end, length, cause and its ticket. Acknowledge an ongoing outage so the team knows someone is on it.
• Shown on status pages — the status pages that use this monitor.
• Recent checks — each check with its result, response time, code and error.

Use Check now to run a check straight away, Pause to stop checking (for example during planned work), Edit or delete. Raw checks are kept for 30 days; the daily figures behind the bars are kept for good.

Status pages

Link a monitor to a component on a status page and the component changes colour by itself, and an incident can open and close on the page automatically.