Appearance
Alert Rules
Use alert rules to identify endpoint conditions that need attention, such as high disk usage or disabled security safeguards. Create them under Library → Alert Rules with Job Designer.
Alerts report conditions; they do not remediate them automatically.
Choose Info, Warning, or Critical severity for any rule. Built-in checks have fixed icons; custom script rules offer a selection of icons. The endpoint sidebar shows active, unacknowledged conditions and distinguishes evaluation errors.
Built-in update and inactivity checks
- OS updates available counts missing updates from Windows Update or Linux APT inventory. Its default threshold is zero and severity is Info.
- Reboot required uses the operating system's reported reboot requirement. An unknown reboot status is not a healthy result.
- Docker container updates available counts running and stopped containers with known image updates, following configured image tracking policies. Cached images without containers do not count. Pinned or unsupported references and incomplete checks cannot establish that containers are current.
- Endpoint inactive checks elapsed time since last successful contact while disconnected, with a default of 30 days and Warning severity. It supports approved endpoint kinds and excludes devices that have never contacted Ordyn. Create and assign this rule to receive prolonged-inactivity alerts; rules and assignments are not created automatically.
Update and Docker checks accept a maximum inventory age, defaulting to 24 hours. Docker registry evidence has a separate maximum age, defaulting to 12 hours. Reboot checks require system status collected within six hours. Choose freshness limits appropriate to your inventory schedules.
Rule sources
Internal
Choose a built-in check when the required value is part of Ordyn's inventory. Configure a threshold, severity, and evaluation interval appropriate to the condition. For drive usage, specify the drive or mount, such as C: or /.
Inventory-based checks need current inventory. Missing or stale observations produce an error rather than a healthy result.
To monitor checkpoints on a guest VM, use Hyper-V checkpoint age too high. See Hyper-V checkpoint age alerts for setup and collection requirements.
Script
Use an endpoint or Runner script with JSON output for a custom condition. Select the value path, its type, a comparator, and the threshold. For example:
json
{
"service": {
"pending_requests": 94
}
}A numeric rule on service.pending_requests with above 90 triggers for that output. The path must exist and have the selected type. A failed script or invalid output produces a rule error.
Endpoint scripts run on the managed device and must match its platform. Runner scripts run on a Windows or Linux Runner and can query an external API or service using the endpoint's stored information. Choose the Runner platform required by the script, independently of the endpoint's platform.
For Runner scripts, configure script parameters on the alert rule. Templates and secret variables use the endpoint's values and overrides. Assign a compatible Runner pool to the endpoint.
Assignments
Assign rules from the tenant, folder, group, or endpoint that should receive them. Review the effective rules on the endpoint's Alerts tab.
A rule applies while at least one assignment covers the endpoint. Removing its final assignment, including leaving a group, clears its current alert and error indicators while preserving evaluation history. Assigning the rule again starts fresh: a new check determines its current state.
Evaluation
Endpoint scripts and inventory-based built-in checks require an approved Agent endpoint. Hyper-V checkpoint checks use host inventory and can evaluate disconnected guests. Inactivity checks support approved endpoint kinds while disconnected. Other built-in checks and endpoint scripts require a connection.
Runner checks support every approved endpoint kind, including offline endpoints, and require an available compatible Runner. Checks can wait while the device or Runner is busy. Review the last evaluation time before relying on a result.
States
| State | Meaning |
|---|---|
| Healthy | The check succeeded without finding the configured condition. |
| Triggered | The condition needs attention at the selected severity. |
| Error | The latest check failed; correct the input or collection problem. |
A failed check does not clear an active condition or its acknowledgement. While the rule applies to the endpoint, the condition remains active until a successful check reports it healthy.
Inspect and test checks
Use an endpoint's Alerts tab to review its rules, assignment sources, and evaluation history. Select an evaluation to inspect the condition and output recorded for that check.
Use Check now to test an assigned rule. The same connection requirements apply as for automatic checks. Manual checks update the alert state and work even when automatic checks are disabled.
Disable checks on one endpoint
Use Disable to pause automatic checks while keeping existing alerts visible. Enable resumes checks if the rule is globally enabled.
Use Unassign to remove rules assigned directly to the endpoint. For inherited rules, follow the assignment source to change the tenant, folder, or group configuration.
Alert automations
An Alert activated automation can start a job when a selected rule becomes triggered. It runs on activation, not on every repeated check. Review the job's target scope and effects before enabling automated remediation.
Dynamic groups
Use alert attention in dynamic groups to organize devices with unacknowledged alerts or rule errors. Acknowledging an alert can remove it from these groups, so avoid treating membership as a permanent record of the underlying condition.