Do you know that most IT leaders use an average of nine—yes nine—different monitoring tools within their environment? That means nine different boxes and nine different applications to manage. And the most interesting part? Very rarely do they even talk to each other. And, this number can be even higher if we bring in hybrid and public cloud scenarios.
Are you managing IT or is IT managing you? Download our eBook to learn more about transforming the IT function.
Managing nine different tools is not only cumbersome, it’s incredibly time-consuming—we’re left reacting to beeps and alerts instead of strategizing how to transform IT. For us to really transform and provide value back to the business, a few things need to change.
First, we need to have a more efficient grasp on business impact. Most tools for IT monitoring are geared toward element monitoring, which is a very granular approach, and only focuses on device level. When you get an alert, you can see what device has the issue, but you don’t know what impact it will have on your organization. By the time you’ve traced the event back to see the impact, you will already be getting alerts directly from your users.
We need to strive for full visibility of business functions to get a clear picture of what’s happening. A technique we use is dependency trees—spending time up front mapping IT elements to the services they provide. The value is that when we get an alert, we instantly know whether it is signal or noise and what the impact it will have, so we can start resolving the issue immediately. As Ben Franklin so notably said, “an ounce of prevention is worth a pound of cure.” In more practical terms, the business doesn’t care that a switch or a VPN connection failed, they want to know what users, locations, or applications are impacted and how quickly are they going to be restored.
Second, we need to eliminate most of the noise we get out of these systems. Noise is the root of all evil in my world because noise means a negative impact in time, effort, and money. For example, if your network links multiple sites and one of those links temporarily goes down it will generate hundreds of events, despite just being one incident. When the link comes back up after 30 seconds, it will generate hundreds more events.
Many organizations still process these events manually—it’s time consuming and runs the risk of blinding your IT staff to critical issues. Think about using techniques like dependency trees and build some automated scripts, and I guarantee you will start to eliminate noise. Depending on the environment, I have seen some organizations eliminate up to 90 percent of their noise. Case in point, Carousel was one of them!
Ready to embark on your business transformation journey by outsourcing run tasks? Inquire today about Carousel’s award-winning managed services.
Lastly, we need to improve our contextual knowledge to provide better and more efficient services; this can come in all shapes and sizes. The quickest and most effective way to do this is to use a “single pain of glass/1-click away method”. After we build dependency trees and remove the noise, if we can give our IT leaders the complete visibility of the eco-system, we also need to create 1-click information for triage and troubleshooting. Too often we need to chase password, IP addresses, circuit IDs in some spreadsheet or SharePoint site. Again, that’s wasted time, effort, and money. Developing monitoring solutions to get 1-click access will drive more efficiency and effectiveness.
We all must start thinking differently as we move forward and how what we do today is helping build a stronger business in the future. For instance, contextual awareness/feeds are going to become very important with DNA/SDN in the future. One of our monitoring partners is already in co-development with Cisco and APIC-EM for different use cases and one of them will be “on the fly” quality of service (QoS).
Think about an executive that needs to collaborate with someone in the warehouse; they will probably have different QoS profiles. But in the future, there will be awareness of both users and build a new profile for the warehouse worker during the call and then tear it down when the call terminates. This will all be done automatically and will require different feeds from monitoring, including Netflow information, device types, wired or wireless connectivity, etc. This is only one small use case in the importance of contextual awareness as we move forward.
In conclusion, monitoring is no longer just an up and down function. We all need to plan for the future and monitoring solutions are paramount to ensure business growth and transformation.