alert

---

cmr/alert

Package: cmr
Type: Directory

An alert rule applies conditions to a set of devices, selected with labels, and runs its actions on the CMR controller when a condition becomes true. A rule fires when a device starts matching its conditions and fires again only after the device stops matching and then matches again. See CMR - Alert rules for how alert rules work.

Flag	Name	Description
X	disabled	Alert rule is disabled.

Argument	Type	Description
name ( unset )	string	Name of the alert rule.
labels ( unset )	object	Select the devices, by labels, that the alert rule applies to. Supports + and - signs as AND and AND NOT operators, respectively; if no sign is provided, the OR operator is used. When labels is not set, the rule applies to all connected devices (labels=all).
connected ( unset )	bool	Alert when the device connects or disconnects. Default: no.
cpu-above ( unset )	num	Alert when CPU load is above the given percent value (0..100); a value equal to the threshold also fires the rule. Default: unset.
cpu-below ( unset )	num	Alert when CPU load is below the given percent value (0..100); it fires only for values strictly below the threshold. Default: unset.
hdd-above ( unset )	num	Alert when disk (HDD) load is above the given percent value (0..100); a value equal to the threshold also fires the rule. Default: unset.
hdd-below ( unset )	num	Alert when disk (HDD) load is below the given percent value (0..100); it fires only for values strictly below the threshold. Default: unset.
mem-above ( unset )	num	Alert when RAM load is above the given percent value (0..100); a value equal to the threshold also fires the rule. Default: unset.
mem-below ( unset )	num	Alert when RAM load is below the given percent value (0..100); it fires only for values strictly below the threshold. Default: unset.
health-above ( unset )	num	Alert when the health sensor named by health-value goes above the given value; a value equal to the threshold also fires the rule. Devices that do not report the sensor are ignored. Default: unset.
health-below ( unset )	num	Alert when the health sensor named by health-value goes below the given value; it fires only for values strictly below the threshold. Devices that do not report the sensor are ignored. Default: unset.
health-value ( unset )	string	Name of the health sensor to compare, matching a name from /system/health/print, for example cpu-temperature. Uses the [cpu-temperature] style placeholder in the action text.
interface-change ( unset )	alt { change: enum (any) , change: ubit (running, not-running, added, removed) }	Alert on an interface state change: any (default) - Any state change. running - An interface's link comes up. not-running - An interface's link goes down. added - An interface is first seen. removed - An interface is no longer seen. running and not-running follow the link state of the interface, not its admin state: enabling or disabling an interface with no connection does not fire, while an actual link loss or recovery does.
interface-type ( unset )	alt { type: enum (any) , type: ubit (ethernet, wifi, bridge) }	Restrict interface-change alerts to one interface type: any (default) - Interfaces of any type. ethernet - Ethernet ports. wifi - Wireless interfaces. bridge - Bridge interfaces.
upgrade-available ( unset )	bool	Alert when a newer RouterOS version is available for the device. Default: no.
disconnected-more-than ( unset )	time	Alert when a device has been disconnected for longer than the given time. Default: unset.
upgrade-job-done ( unset )	alt { cond: enum (yes) , cond: ubit (success, fail) }	Alert when an upgrade job finishes; the rule fires once for the whole job, when the last device of the job is done: yes (default) - Any outcome. success - The job finished successfully. fail - The job failed. A job that ends without installing an upgrade is treated as a failed job, so the yes and fail variants both fire.
upgrade-done ( unset )	alt { cond: enum (yes) , cond: ubit (success, fail) }	Alert when the upgrade of a device finishes; the rule fires once for every device that finished its upgrade: yes (default) - Any outcome. success - The upgrade finished successfully. fail - The upgrade failed. When no newer version is available for a device, its upgrade still finishes as a failure (the [upgrade-error] is no upgrade available), so the yes and fail variants both fire.
unpaired-device-connected ( unset )	bool	Alert when a device that is not paired to the controller connects. Default: no.
rebooted ( unset )	bool	Alert when the device reboots. Default: no.
log-topics ( unset )	multi { array-id, array-id, topic: super { ! , topic: enum } }	Log topics to follow; prefix a topic with ! to exclude it.
log-regex ( unset )	string	Alert on log lines the device sends to the controller whose message matches the given regular expression.
action.log ( unset )	string	Write a message to the log when the alert fires. The message may contain placeholders such as [device] or [cpu-usage], replaced with the values of the device that fired; an inapplicable placeholder is rendered as unknown.
action.log-topics ( unset )	multi { array-id, topic: enum }	Log topics to write the message to. Default: cmr,info.
action.log-prefix ( unset )	string	Prefix for the log message. When set, the prefix is prepended to the message without a separator; when unset the message is written as it is.
action.script ( unset )	enum	Run a named system script when the alert fires. The script must allow itself to run with reduced rights; otherwise the action fails with executing script ... from cmr failed ... (not enough permissions). Set the script's dont-require-permissions to yes.
action.script-vars ( unset )	multi { array-id, script-var: string }	List of variable names to make available to the action.script: each variable is filled with the value for the device that fired. For example device, identity, address, board, version become $device, $identity, and so on in the script. A name with a hyphen (for example cpu-usage) cannot be read as an ordinary script variable.
action.http-url ( unset )	string	HTTP webhook URL to call when the alert fires. On a failed call the alert writes alert "<name>" HTTP action failed: <reason> to the cmr topic with warning severity and counts the failure in action-failures; the other actions still run.
action.http-method ( unset )	enum (get | post | put | delete | head | patch)	HTTP method used for the webhook: get post put delete head patch
action.http-body ( unset )	string	HTTP request body sent to the webhook. Supports the same placeholders as the log message, substituted per device (for example [device] and [version]). No Content-Type header is added automatically; add one with action.http-headers when the receiver expects it, for example Content-Type: application/json for a JSON body.
action.http-headers ( unset )	multi { header: string }	Additional HTTP headers sent with the webhook, each as Header: value. The request also carries the controller's RouterOS <version> user-agent.
reset-on-disconnect ( unset )	enum (yes | no) { yes:0, no:-1 }	Clear (reset) the alert state when the device disconnects, so a reconnecting device fires the rule again: yes or no. Default: no (the fired state is kept across a disconnect).
severity ( unset )	enum (critical | high | medium | low)	Severity of the alert: critical high medium (default) low
category ( unset )	multi { array-id, category: enum }	Categories assigned to the alert. The category is a grouping label: the server assigns one according to the rule's conditions (for example performance, availability, version, network, or logging), and it can be changed to any value.

Read-only Argument	Type	Description
devices	num	Number of devices the alert rule matches (the devices selected by labels). When the rule covers all devices, this count includes the controller itself.
devices-on	num	Number of devices the alert is currently fired on.
fired	num	Total number of times the alert has fired.
action-failures	num	Number of times an alert action has failed, for example a webhook that could not be reached.