Apr 23

In today’s fast growing e-commerce market, the ability to develop and deploy web applications quickly could make or break a company’s future. However, developing and launching live web applications alone just does not merely guarantee company’s success. Without proper systems in place to monitor, verify and detect symptoms of a problem, management of such applications can soon turn into a manageability crisis. As many organizations adopt web-based applications fueled by latest developments in web 2.0 technologies, the stake for accurate management systems has never been so high.

As organizations quickly embrace various technologies to improve their web applications, various web technologies are mashed to form a single/collage web application. Undoubtedly, the complexity of such an “integrated” application also increases exponentially. As these applications interact with in themselves and end users, they become extremely data intensive. Hence, the availability of data becomes extremely important. As it goes through several testing cycles, various scenarios are played out and the code is debugged before the application goes live. However, when the application goes live, end user scenarios prove to be completely different. Thus, there could be a distinct disconnect between the end user experience versus the use cases tested.

For example, lets us consider British Airways (BA) which lets the end user book tickets through critical live web applications. With worldwide customers, it becomes crucial for BA to be able track their web applications health and availability 24×7x365. Through global Point of Presence, such monitoring and alerting is carried out by finding out the response times of websites from different geographical locations. Several extraneous factors, fuzzy boundary conditions, quality of data input are also likely to affect the performance of the web application. These factors could negatively impact the performance of the web application and in extreme cases - may even derail the entire application. Such a derailment could affect BA’s revenues - which could range from a few thousands of dollars to millions/billions. In addition, the company’s image could also be adversely affected.

In mission-critical live web applications detecting the failure is only the half the work. Finding out the root cause of the failure becomes more important.

 

Scenario in an application monitoring is not present

Figure 1: Scenario 1 – Without proper monitoring tools, failures go un-noticed

Scenario when application monitoring exists

Figure 2: Scenario 2 – With proper monitoring tools, failures are monitored and informed as it happens

 

Methods such as URL monitoring and URL sequence monitors, which are actively practiced by ManageEngine Applications Manager, are used to detect such failures actively. These monitors provide live monitoring facilities, which enable the IT team to track web applications health and availability. Based on certain conditions and parameters, alerts are then issued to the IT team. By providing such active alerts, end user experience measurements help any organization to be able to track their web applications live and help them to maintain the performance of the application at optimal level.

 

Better end user experience = Better Revenues = Better Reputation = Loyal Customers

 

Many enterprises/organizations have started looking at these parameters as it becomes a part of the vital statistics of what the end user want/need. It also helps the IT team to detect any kind of failure occurring within the web application and thus avoiding failures in the future as well. Maintaining better end user experience measurement would lead to better retention rate of a company’s customer base. It also helps in acquiring new customer base and ultimately translates into better revenues and market share growth.

Apr 22

A bug in our backup script - Don’t worry, the fix would be out in our next Service Pack. But we a owe a big thanks to Brett Peterson, VisionShare, Inc. He not only pointed out to the bug but also suggested a way to fix it up.
————
The bin/BackupDB.sh script contains a bug that prevents it from running to completion at any time in the A.M. The SQL dump filename is generated using the following command:

export name=`date +%F_%k_%M_%S`

The %k causes a space in the resulting filename if the hour is not two digits in length (i.e., anytime in the A.M.). This causes the MysqlZip invocation later in the script to not find the .sql file and error out.

I patched this by changing the command above to use %H instead of %k:

export name=`date +%F_%H_%M_%S`

This causes a two-digit A.M. hour to have a zero in the first position and thus no spaces in the file name.

————

Susan

AdventNet.com | ManageEngine.com