In this series of blog posts, I would like to identify areas in IT operations where fusing AI will be compelling to do. Before we decide that AI should be fused in every day’s IT operational activities, we should present traditional (non-AI) IT operational use cases and highlight corresponding pain points. This will highlight the need to introduce innovations to address the pain points. The perspective presented here is product-agnostic and is based on many customer engagements that I had to lead for more than a decade.
What I am going to address here are the activities that will take place after the software solution is deployed in production. Below are the major activities that take place after the solution is deployed in production. Additional activities that you may have seen elsewhere can be included as sub-activities of the below activities. In this blog post, I will only provide a brief introduction to each activity. More details for each activity will be provided in the near future.
- Identify
In this activity, the solution owner needs to identify:
- The key performance indicators (KPIs) that need to be collected across all solution layers.
- The threshold values to set for the various KPIs. Once the threshold value is reached, an alert is generated.
This activity is not done just once. The solution owner will often identify additional KPIs and set their thresholds to track as long as the solution is running in production.
- Monitor
In this activity, the solution owner needs to identify and implement a solution to monitor the key performance indicators that have been identified in the previous activity. Once the solution is identified and implemented, it runs constantly to track the KPIs identified above.
- Alert
In this activity, the solution that the solution owner has implemented in the previous activity generates an alert when a KPI threshold is reached.
- Resolve
In this activity, the solution owner receiving the alert resolves the problem leading to the alert.
- Notify
In this activity, the right person involved in the solution is notified. This activity happens often during the alert activity as well as the resolve activity. When the alert happens, the right people must be notified. The Notify activity also happens throughout the resolve activity.
The degree of maturity of production operations varies from one enterprise to another. Believe it or not, I was leading the optimization of a production environment for an enterprise where a complex solution was deployed in production without proper operational procedures.
Do you have the right operational procedures for your production environment? In the next blog posts, I will take a deeper dive into each activity and highlight pain points that will drive the need for fusing AI into operational procedures of production environments.