An alert watches one measurement on a node, such as processor load or free disk space. When the measurement crosses a limit you set, the alert notifies the people you choose and leaves the node as it is. Set one up for whatever you want to hear about first: a disk filling up, memory running short, or a website that stops answering.
Before you begin
Section titled “Before you begin”You need:
- A node in the Running state. Create Alert asks for it, since only a running node sends monitoring data to alert on.
- Monitoring data on the node’s details page. If its graphs are blank, fix that first; see Monitoring graphs are blank.
- An alert group, the list of people an alert notifies. Alert groups are kept in Event Monitoring; see Manage alert groups. You can also create one from the alert form.
Create the alert
Section titled “Create the alert”- Open the node and go to its Alerts tab.
- Click Create Alert. The Create Alert form opens.
- In Alert trigger type, choose what to watch:
- Processor Load: how busy the node’s vCPUs are, as a percentage.
- Consumed Memory: how much of the node’s memory is in use, as a percentage.
- Free Disk Space: how much space is left on a disk, as a percentage.
- Volume Read Ops and Volume Write Ops: reads or writes on one disk.
- Webcheck: whether a web page still answers with the text you expect.
- For Free Disk Space, Volume Read Ops, or Volume Write Ops, choose the disk in Volume. Pick Node Volume(Default) for the node’s own disk, or an attached volume by name.
- Set the condition. For every type except Webcheck:
- Operator: Greater than, Less than, Equal to, or Not equal to. To hear about a disk filling up, watch Free Disk Space with Less than; for the other types, Greater than is the usual choice.
- Threshold value: from 10 to 100 for the three percentage types, and from 1 to 75,000 operations for the two volume types.
- Severity: Not classified, Information, Warning, Average, High, or Disaster. For what each level means, see Severity.
- For Webcheck, fill in these instead:
- Monitor URL (HTTP 200): the address of the page to check, such as
https://example.com/health. The page is expected to answer with HTTP status 200, the code a web server sends when a page loads normally. - Webcheck Interval: pick one of the listed values.
- Contain Content/Text: text that is always on the page when it works, such as
OK. The alert fires when the page does not contain it.
- Monitor URL (HTTP 200): the address of the page to check, such as
- In Alert Group(s), choose who to notify. All picks every group. To make a new group, click the plus icon beside Alert Group(s). The Create New Alert Group panel opens, and you stay in the alert form. Create needs at least one alert group, so make one first if the project has none.
- Read the Alert Summary line. It describes the alert in one sentence, so you can catch a wrong operator or threshold before you save.
- Click Create.
A node can have up to five Webcheck alerts, each with its own combination of URL and text.
Check the alert on the Alerts tab
Section titled “Check the alert on the Alerts tab”The alert appears on the node’s Alerts tab straight away, with its condition written out as a sentence, for example:
Alert when average load is greater than 80% of vCPUA Webcheck alert reads “Alert when Web server URL” followed by the address and the text it looks for.
Notifications go out by email to everyone in the alert groups you chose. The account owner can also get them on WhatsApp; see How alerts reach people.
Pause, resume, or delete an alert
Section titled “Pause, resume, or delete an alert”On the Alerts tab, each alert has a pause icon and a delete icon. Each asks you to confirm before it acts.
- Pause Alert stops the alert’s notifications but keeps it in the list, for example during planned maintenance. While an alert is paused, it offers Resume Alert in its place.
- Resume Alert starts the notifications again.
- The delete icon removes the alert from the node.
To change an alert’s settings, use the Monitoring Dashboard link on the same tab. It opens Manage Alerts, where you can edit the alert; see Edit, pause, or delete an alert.
If a notification does not arrive, see Troubleshooting node monitoring and encryption.