B-3-18CRYF9

Sizing your Cloud Monitor Data

Blog Post created by B-3-18CRYF9 Employee on Oct 7, 2015

Cloud Monitor is a real time logging API designed to allow you to easily instrument and monitor all transactions that are delivered through Akamai, and deliver the log data to one or more destinations in a standardized format.  The logs can be used for any number of BigData queries and is of enormous interest to customers.  Today, customers are charged based on number of log lines. One of the questions that customers ask is how do I size the log information?  Customers need this for

  1. 1.              Budgeting purposes.
  2. 2.              Estimating their storage costs.

This note is aimed at providing some easy answers.  In the example below, I will be using www.akamai.com as the customer.

Cloud Monitor generates a log line for every edge hit that we encounter for the specific customer.  The Luna Control Panel provides this information.  Selecting User Traffic from Monitor menu option gives you the following screen.  Notice the Edge hits has been circled. This will be for a pre-defined set of CP Codes and for the last 2 days.

Picture1.png

 

Next look at what this number of edge hits represent.  It may not represent the CP Codes that the customers would want.  By selecting the ‘Change’ menu option, you can select the CP Codes and the time period that would be of interest to the customer.

 

Picture2.png

 

For the purposes of this illustration, I choose all the CP Codes.  You then change the Date Range to ‘1 month’.

 

Picture3.png

 

For www.akamai.com this is what I get (as of July 24th 2015, 8 AM):

 

This means for the last 30 days, if Cloud Monitor was enabled, 2,749,273,063 log lines would have been delivered.  I like to be safe and would suggest to the customer that they saw approximately 3 billion edge hits last month, this would mean that would have had 3 billion log lines delivered.

 

The next question the customer is going to ask is, how does this translate to my data-store size.  I generally do not get into database sizing – this varies widely based on several factors.  Instead, I would give them an estimate of how big this represents in raw data.

 

A typical log line averages 1KB today.  Using this 3 billion log line example, for www.akamai.com, the data that would be sent to the customer would be approximately 3 TB (terabytes).

 

Cloud Monitor does provide the capability to add custom data, cookies etc. If these are added, the amount of data delivered would go up and the amount  This exercise is based on no additional data being added.

 

As always, if there are questions, please use your internal Akamai resources.

Outcomes