{% extends "base.html.twig" %} {% block title %}How to manage Alerts{% endblock %} {% block body %}

How to manage Alerts


Alerts in general

Alerts purpose is to monitor the defined criteria (e.g. device temperature, availability, free disk space, number of reboots) and trigger, siwtching to the Raised state when criteria are met. Based on notification settings, a notification can be send via one or more notification channels as well.

List of alerts is sortable and searchable, similar to the devices listing.

The table shows some of the alert definitions, but it takes some time to learn how to read it. Before getting to the definitions, note these columns:

  • Triggered ... Shows if the alert is triggered at the moment.
  • Triggered time ... Shows time of the most recent allert triggering.
  • Count ... Shows how many times was the alert triggered.
  • Enabled ... Tells if alert is enabled (default) or not. This way you can deactivate any alert.
  • Active ... Tells if the alert is actually being checked during alert evaluation. In case there is an error in definition, alert may become non-active (also called Invalid). Also, there is a brief moment the alert stays Inactive just after alert definition change and before the system picks up the new definition and validates it, setting the alert to Active.

Notification channels

You need to setup notification channels in System -> Settings so that notification can be delivered to you once alert triggers.
Currently supported notification channels are:

  • Slack
  • E-mail - an opt-in channel. Should you want to get e-mails, you need to check Enable Email Notification checkbox and provide a valid e-mail address.

Once you setup your channels, notification for any triggered alert is sent to all enabled notification channels.

Alert target

Alert can check one target type. The most used target types are:

  • Device ... Alert is checking a particular device (e.g. monitoring one particular device availability).
  • Number (#) of devices in a group ... Alert is checking all devices in a group (e.g. monitoring temperature of every device in a group) and if the number of devices matching the condition matches or exceeds the defined number, the alert triggers (e.g. trigger when at least one device in a group X temperature goes out of range 20 .. 60 degrees of celsius).

Defining alert

To define an alert, you need to click on Create Alert first. Once on the alert creation page, you should select:

  • Name ... alert name is used in notification so you should select something recognizable enough to tell you in the notification which alert was triggered.
  • Target type ... Select if you want to monitor a particular device or a number of devices in a group.
    • If you select a device, you need to select particular device by DevId or device name (if you are naming you devices)
    • If you select to monitor a number of devices in a group, you need to select a group and a number/amount of devices (in the sense of at least a # (number) of devices in this group...
  • Target specification ... here you select what characteristic/metric are you monitoring. (e.g. device temperature)
  • Operator ... here you select logical operation which is checked against one or more values mentioned below. (e.g. outside of)
  • First Number ... enter/select first threshold (enter a number, e.g. 20 (as in 20 degrees))
  • Second Number ... enter/select second threshold - if applicable - (enter a number, e.g. 60 (as in 60 degrees))

Example alert

The following alert triggers whenever a device in DWARFG_LONG system fails to contact server for 5 or more minutes.

  • Click on Create Alert button
  • Name ... enter text, e.g. Device lingering
  • Target type ... select at least # devs in grp
  • Target monitoring group ... select the 'All devices' group: All devices
  • Target amount ... enter 1 so that a single device matching condition can trigger the alert: 1
  • Target spec ... select Data Age
  • Oper first ... select greater than
  • Num first ... select 5 minutes
  • Click on Create Alert confirmation button

Note that the alert is listed as Enabled but Inactive. It takes a few seconds for alert validation before switching to the Active state. Please reload the page after a short while. If you still see the alert as Inactive, there is a problem with alert definition.

{% endblock body %}