Review and run a host's monitoring
Read the status of each monitor and its attempts before alerting, run it on demand, and decide whether it opens an incident.
On this page
The Monitoring tab of the page shows in a single table everything checked on the host: its availability and metrics, the associated tests, the probe checks and the agent plugins.
Steps
Open the host page and go to the Monitoring tab.
Use the filter to narrow the table: All, With problems, Tests, Probe or Agent. You can also type in Search monitoring….
Review each row:
Column | What it shows |
|---|---|
Status | The confirmed status: OK, Warning, Critical, Unknown, Not checked yet or Disabled. If the last result differs from the confirmed one, you'll see · SOFT (see below). |
Name | What is checked, what type it is, who runs it, how often and, if it came from a template, the template's name. |
Last update | How long ago the last result arrived. |
Last change | How long ago the status changed. |
Attempts | Consecutive failures out of the maximum before alerting, for example 2/3. |
Status information | The output of the last result and its performance data. |
Creates incident | Yes if it opens an incident when it fails; No if it's recorded only as an event. Below, the resolver group if it has its own. |
Dependency | The monitor it depends on. If that one is down, this one's failure is recorded only as an event. |
Confirmed statuses and attempts
A monitor doesn't alert on the first failure. It counts consecutive failed attempts and, when it reaches the maximum, confirms the new status (HARD status). Meanwhile, the last result is shown as SOFT: for example, OK · SOFT means the confirmed status is still OK but the last check failed. Hover over the status to see the details.
Run a monitor on demand
In the monitor's row, click Run.
If the cloud runs it, the result appears in a notice right away. If a probe runs it, the notice says it was sent and the result arrives within seconds; meanwhile the button says Running….
Some monitors can't be run on demand. In that case the button appears disabled and, when you hover over it, explains why.
Decide whether it opens an incident
With permission to operate, use the row's Creates incident switch:
Yes: it opens an incident when it fails.
No: it's recorded only as an event when it fails.
The change is saved immediately. To choose its own resolver group or a dependency, click Configure monitoring and use the Monitoring tab of the configuration.
Monitoring from the probe
If a probe checks the host, the Monitoring from the probe panel appears below the table with:
The probe, protocol, port and interval, and the Change button.
Each instruction with its type, interval, last result and output.
Depending on the protocol: Interfaces (SNMP), Processes (top consumers), Windows services, Extra OIDs, HTTP (with the days left on the certificate) and Received from this host (syslog and traps received).
For SSH and administrators: Re-pin SSH fingerprint, for when the host's fingerprint changed because of a reinstall or a new key.