Interval, Sample and QoS Interval meaning in CDM GUI

Products

DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

This KB contains a practical explanation and examples on how CDM Control properties work.

What does Interval, Sample and QoS Interval mean in the CDM configuration window?

How can I control how often CDM sends QOS and alarms

Environment

DX UIM 20.4.* / 23.4.*
CDM any version

Cause

Guidance

Resolution

Here we have 3 Values for each Disk, CPU and Memory properties:

There are three concepts here (Interval, Samples and QOS Interval): cdm IM Configuration

Interval: Means how often the system checks for data. In other words, specifies the interval (in minutes) after which the monitoring information is obtained from the system.
When the CDM probe actually queries the system for CPU, DISK or MEMORY. When data is picked up from a system, is not necessarily inserted to the DB yet, but it's kept in memory and is used for threshold comparison and is inserted in the DB only after the QOS interval value is also met.

Samples: Means how many data points are averaged on a rolling basis. The samples specify the number of samples the probe stores. This number is used to calculate values to compare with thresholds to send alarms. The alarms are sent only if it has enough samples.

Qos Interval: (Multiple of 'Interval'): Means how often the QOS data actually gets written to the database. Specifies the monitoring interval between each sample collection for QoS information. For example, If the interval is set to 5 minutes and a number of samples is set to 5, the average CPU utilization for the last 25 minutes is displayed in the QoS message.
If “the value is greater than 1, the probe only calculates the average data for CPU monitoring. For disk, memory, and paging QoS messages, the probe generates the QoS messages with the current value.

The practical Example below is for CPU properties:

----------------------------------------

In this example, Interval is 5 minutes, Samples is 5, and QoS interval is 2.
Data is written every 10 minutes (Interval x QoS Interval) into the database.
----------------------------------------

This is How QoS gets written to the database:

ie. CPU properties: interval 5 min, samples 5, QoS Interval 2.

# time cpu sample (how number is calculated)

a 1:00:00 50% ( a )
b 1:05:00 51% ( (a+b)/2 ) qos to database
c 1:10:00 50% ( (a+b+c)/3 )
d 1:15:00 50% ( (a+b+c+4)/4) qos to database
e 1:20:00 50% ( (a+b+c+d+e)/5 )
f 1:25:00 75% ( (b+c+d+e+f)/5 ) qos to database
g 1:30:00 75% ( (c+d+e+f+g)/5 )
h 1:35:00 50% ( (d+e+f+g+h)/5 ) qos to database
i 1:40:00 50% ( (e+f+g+h+i)/5 )
j 1:45:00 51% ( (f+g+h+i+j)/5 ) qos to database

This is How Threshold Comparison work for Alarms:

"Samples" affects the averaging applied to the measurements used to compare against thresholds (alarms).
In this example, samples are set to 3, interval set to 5 and QOS interval set to 1.

In the example, CPU Measures correspond to:

90 (%) after 5 minutes
90 (%) after 10 minutes
100 (%) after 15 minutes

The last 3 QOS values (samples) sent after 15 minutes are then values 90, 95, and 100. After you have met the "samples" number of samples, the probe will start comparing with the thresholds and that first comparison will be (90+95+100)/3= 95%.

If after 20 minutes you have another 100% QOS metric your next comparison will be (95+100+100)/3 = 98.3%.

If you are still at 100% after 25 minutes in, your next comparison would be (100+100+100)/3 = 100%.