Smart Alarms

Smart Alarms

Smart Alarms 2560 2304 Kronometrix

What are Kronometrix Smart Alarms?

Kronometrix is a distributed data management fabric that is capable of receiving and processing many different types of data messages from many different data sources. Everything from ICT and IoT to weather stations and environmental devices. At the same time as offering extensive data-source flexibility, the Kronometrix Smart Alarms capability can also send alarms if certain predefined conditions pertain. For example:

  • if the data is delayed and does not arrive on time
  • if the data does not meet certain numerical formats
  • if user-defined business logic is violated

Kronometrix Smart Alarms uses a special algorithm, based on sampling the prescribed frequency and duration of trigger events, to eliminate the typical flood of unnecessarily repetitive alarms; the bane of every on-call technical support person.

For example, the Kronometrix operations team was originally receiving around 500 alarms per day. That level of alarms becomes highly irritating noise that eventually tends to gets ignored (thus defeating the purpose of the alarm). After implementing the new Smart Alarms interface, that number was dramatically reduced to less than 20 per day.

All Smart Alarms are standard feature of K50, K500 and Kronometrix SaaS.

How does it work ?

For user based alarms, these are the main important concepts used:

  • Threshold
  • Frequency
  • Duration
Threshold

The Threshold is a numerical value used to define a boolean condition: true or false. When that boolean is true, a count of that event type is accrued toward triggering an alarm in the future, depending on the settings of the Duration and Frequency parameters.

Frequency

The frequency is the rate at which events need to occur in order for an alarm to be issued. This rate is defined through two quantities: the number of events and the timebase. For example: to specify that an alarm be issued when three events occur in one hour, the user would enter a 3 for the number of events, and 3600 (seconds) for the timebase.

Screenshot from Kronometrix Smart Alarms User Interface

Duration

The event duration is defined as the period during which a set of contiguous data samples meet the user-defined threshold condition. An OVERRIDE value sets the maximum number of alarms that will be issued. Once this number of alarms has been issued, no further alarms will be sent, even if the alarm condition is met in the future. The default value is zero and means the OVERRIDE is disabled so that alarms will always be sent when the alarm condition is met.

Examples

1. Send continuous Smart Alarms as soon as the CPU utilization on the database server exceeds 75%, AND there are at least 5 such events, each lasting more than 300 seconds during a moving time-window of 1 hour. If the same situation arises some time in the future, send another Smart Alarm. Otherwise, don’t send any more Smart Alarms. The user-defined settings would be:

  • Number of events: 5
  • Timebase: 3600 seconds
  • Event duration: 300 seconds
  • Override Count: 0, default

In this example, Smart Alarms will send continuous alarms as soon as the conditions are satisfied in each occurrence.

2. Inform me if our service-desk application becomes really slow. Send just a single Smart Alarm if the Response time of the application exceeds 500 ms over a 15 minute period, and there are at least 5 such events in minute. The user-defined settings would be:

  • Number of events: 5
  • Timebase: 900 seconds
  • Event duration: 60 seconds
  • Override Count: 1

In this example, Smart Alarms will send a single alarm during the period when the condition is true.

3. Send me the earliest single Smart Alarm when the air temperature falls below 0 ℃, just once, in a 20 minute window. Beyond that, dont inform me anymore unless the situation has changed. The user-defined settings would be:

  • Number of events: 1
  • Timebase: 1200 seconds
  • Event duration: 1200 seconds
  • Override Count: 1

In this example, Smart Alarms will send just a single alarm until the temperature condition is no longer true.

Leave a Reply

Back to top