Efficiently detecting and resolving problems is essential, of course, to continue supporting - and minimizing impact on - business services, as well as minimizing any financial impacts.
The goal is to turn the tables on IT problems so that 80 percent of the time is spent on the root cause analysis versus 20 percent on the actual problem fixing.
In resolving the issue, communication is a critical factor for integrating different expert groups towards a common goal. Because each team holds a narrow view of its own domain and expertise, there is always the danger lurking that the "big picture" angle will be missing. You don't want lack of communication to result in blame games and finger pointing.
Some problem detection methods include:
- Infrastructure Monitoring: specific resource utilization like disk, memory, CPU are effective for identifying availability failures – sometimes even heading those off before they happen.
- Domain or Application Tools: These help, but leave the issue that overall problem detection is still a game of hide-and-seek, a manually-intensive effort that comes under the pressure of needing a fix as quickly as possible.
- Dependency mapping tools, which map business services and applications to infrastructure components, can help you generate a topology map that will improve your root cause analysis process for the following reasons:
1. Connect Symptoms to Problems: A single map that relates a business service (user point of view) to its configuration items, will help you detect problems faster.
2. Common Ground: The map ties in all elements so that different groups can focus on a cross-domain effort.
3. High-Level, Cross-Domain View: Teams can view problems not only in the context of their domain, but in a wider view of all network components. For example, a database administrator analyzing a slow database performance problem can examine the topology map to see the effect of networking components on the database.
Root cause is a complex issue, so that no single tool or approach will provide you with full coverage. The idea is to plan a portfolio of tools that together deliver the most impact for your organization.
For instance, if you do not have a central event management console, then consider implementing a topology-based event management solution. If most of your applications involve online transactions, try to look for a transaction management product that covers the technology stack that is common in your environment. Put differently, select a combination of tools that are right for your environment.
Once you assess the tools that provide the most value, implement them in ascending order of value so that you get the biggest impact first.
Ariel Gordon is VP Products and Co-Founder of Neebula.
