AIOps for applications

Author: Mr. Nagarajan Pichumani

Published on: March 31, 2018

The recent Gartner report states one of the features introduced in the modern day APMs, as follows:

"Artificial intelligence for IT operations (AIOps) for applications — AIOps for applications enables the automated discovery of performance and event patterns, and detection of the source (or root cause) of performance anomalies for HTTP/S transactions supported by Java and .NET application servers. This is accomplished through machine learning, statistical inference and/or other methods. In 2016, AIOps for applications was referred to as "application analytics" and then as "algorithmic IT operations," before being changed to "AIOps" in mid-2017".

In today's world, use of the terms machine learning or AI is very important from branding perspective. Whether the customer uses ML or AI, they want to talk on these lines! But from our interactions with at least 100 customers so far in the last 6 months, most of them are not even using the regular business intelligence practices! When we dug deep on why the regular business analytics practices/tools are not followed in their core business, we consistently got the following top 2 answers:

      1. 20-25% of the data collected is not clean - at least they are not sure on the data cleanliness
      2. User community of the business products are so used to MIS text reports, and not ready to change to visual analytics

First of all, the application performance data thru APM tools is not generated by humans, but by the APM agents from the machines and applications. Hence there is no question of the data being unclean. So the first part is resolved.

The IT infra team holds a key position for the business apps to run. Hence ensuring the speed and availability of those systems is extremely important. Once we provide the visual analytics, both real time and historic, the IT team will be the first customer within the organization. From APM tools itself, a set of business analytics can be done easily..

For example, the number of ATM transactions posted, number of shopping carts abandoned and/or purchased, number of payment gateway timeouts etc. can be obtained thru APM tools. This is achieved by collecting metrics on specific pages and backend routines. When IT team starts providing these critical business metrics, we see a natural pull by the senior management to ask for more business analytics.

Moral of the story: Automate IT Ops management with APM tools, collect business metrics thru that; very shortly your business team will embrace business analytics.

You May Also Like


What about the business performance?


Be careful with smaller enemies