Every few days I hear stories in the press, on social media and by talking with people about the major outages that happen to large enterprise applications. The implications for companies that suffer a major disruption span from embarrassment to significant lost business to huge costs and all the way through to a loss of reputation or a devaluation of stock prices. And are often complimented by regulatory and compliance issues leading to further investigations and even fines.
What you don’t hear about when a major system outage happens are slaps on the back for a great job, promotions and bonuses being paid.
And that’s the difference between monitoring, and tracking transactions! Let me explain:
Monitoring is the process of continually collecting data points on each system. Normally this includes how long something takes, or how much resource it uses to do its work. The challenge is that monitoring doesn’t often tell you if something is producing the right result, just that it produced something. If a process provides the wrong result, it can do this without using a lot of resource and may do it quickly. Companies that rely on the classic monitoring approach, then spend inordinate amounts of time and money building complex scripts that help them understand the gap between a heartbeat (it’s alive), and the quality of the results.
Tracking a transaction is the process of looking at the data that is input into a process and the result that are issued, and it will include the timing detail about how long it took as well as what happened inside the process. If the process issues an invalid response this can be seen. Tracking a transaction through every system that makes up a business process allows a business to see exactly how the business performs, and the way in which every customer experience’s it.
The difference can be profound. Those who track transactions can see business issues forming well before the user’s experience is impacted, while those using just classic monitoring can only see a limited view of the technical data, and only through immense scripting effort can they define the relationships well enough so they start to see business performance issues. Even then they will not be able to investigate any alerts quickly or effectively enough to be able to define the root cause.
To track transactions demands knowledge of the infrastructure, platform, applications and middleware. That’s what it takes to move from a reactive operations approach where false positives can make command and control frenetic to a pro-active operations approach where data is turned into information and then into knowledge that increases performance, increases customer satisfaction and reduces costs.
There are times when just monitoring systems is enough, but for large enterprise applications that demand maximum efficiency and effectiveness you should consider tracking complete transactions.
Find out more at www.nastel.com