Complexities of Business Intelligence and Analytics ecosystems

Business Intelligence and Analytics tools are complex systems of infrastructure. Not only do you have to keep the core tools running, you also need to insure all of the supporting infrastructure like databases, ETL processes, firewalls, logins, browser settings and more are all working.

"Kill the messenger" is a fact of life and typically, the BI solution is the messenger. Even though the problem lives somewhere else, you will be blamed for it.

There are different types of failures you can experience and all will impact your output. Any one of them failing causes output problems, some minor and some major.

In following posts, I will cover some things you can do to fortify and monitor your environment against these failures to provide better service even in the light of the failures being beyond your responsibility area.

Here's a partial list of your BI/Analytics supporting domains and some failures that can reflect on you:

Source databases

If the source of your data has a problem, you will not be able to obtain data for your reporting stores and thus be presenting stale data. This problem could be permissions, password expiration, outages, firewall changes, unannounced database changes and more.

Data
Errors in your source systems, say a UI or other bug in the actual application can cause cause bad data to be generated and you to pick up said bad data.

Data Transformation
Your ETL or other data transformation processes could fail or worse yet, do its job wrong and create bad data. It could terminate early, be locked out from its' source database and more.

Delivery Systems
Your reports could execute perfectly but perhaps your Email system is down. Maybe your report suddenly grew in size to exceed attachment limitations and fail. Perhaps you FTP the output to another system and a firewall change stopped your delivery.

Communicating these issues to your stakeholders on a regular basis is the best advice I can give. You want them to understand as much as possible that you are only responsible for part of the whole solution and thus subject to organizational dynamics, policies and other things that will delay your return to healthy reporting BI and Analytics.