Let us see if we can find the five leading reasons (maybe more, maybe less) for why we need a proactive or predictive solution these days.
#1: I don’t have effective change control in place that spans into and incorporates the monitoring that I do on end point systems, applications and services.
#2: My boss wants me to “do more with less” so I need to figure out a way to clean up the mess I have today in my resource monitoring and event management solution.
#3: I know that when this thingy begins to slow down and that thingy drops packets that my transactions begin to fail. Now how do I write that policy to correlate all my thingys?
#4: My tool is better than your tool. I need to figure out a way to make you believe that your tool is always wrong so you’ll work my trouble ticket.
#5: My manager told us that we need to become more proactive. I sent the dba an email to tell him that we were going to have an outage to this database in three hours. He’d already gone home for the day.
These are tongue in cheek, but the underlying themes of each one are very valid in nearly all operations and application support groups. Why are we interested in predictive and proactive tools when we probably don’t have our own house in order in the first place?
How would you write the business justification and capital purchase plan to explain why you need them? How will you quantify your reasoning? Are you willing to give up one or more FTEs to purchase this solution? Have you had an honest look into the far reaching corners of your organization to see where the real root causes may be that spark your interest in these solutions? Are you ‘really’ ready to try and be proactive or predictive? Are you ‘really’ doing reactive well? What does predictive and proactive really mean to you? How would you describe the core capabilities such a solution should have? How would you associate expected value and ROI from having those capabilities? Where should we be looking elsewhere for help in these areas (BI, operational BI, BPM, BAM, analytic databases, statistical modeling and forecasting, etc.)
Please share your thoughts and ideas on why proactive and predictive solutions are of interest these days.
Comments on this entry are closed.
Doug, excellent post!
Interesting questions, Doug. It would be great to hear from some customers of Real time analytics solutions (proactive/predictive solutions) here, however there is one important thing I’d like to point out.
Although the “sexy” part of the real time analytics solution’s capabilities is the proactive/predictive stuff, there are other capabilities that can make a big impact for an Ops team. For example, simply understanding the normal behavior of your infrastructure can significantly improve the problem resolution process. It automates a lot of the manual work that goes into sifting through siloed monitoring alert storms trying to figure out which alerts are relevant to the current problem and which are not. It eliminates false positives related to normal behavior obscured by static thresholds and uncovers real problem precursor behavior. These solutions can also provide an understanding of how infrastructure components are related to user experience and business performance. This information can be used to build an adaptive Performance Management Database (PMDB) that helps identify performance relationships between IT infrastructure components much like a CMDB defines their physical relationships. You can get reports that show you the top infrastructure influencers of business performance and user experience key indicator (KI) breaches. Rather than using tribal knowledge to address problems, you have a map that indicates what to look for based on actual performance. Additionally, capturing models of the abnormal behaviors that lead to KI breaches provides a forensics capability that can pinpoint root cause of problems much more efficiently tha getting a representative of every silo on a bridge call. None of the things I’ve mentioned here are about prediction or proactive alerting, these capabilities are about providing the intelligence that fuels improved troubleshooting efforts. The focus in the press and analyst community on the nirvana of predictive alerting (and the challenges) sometimes diverts attention from other capabilities of real time analytics solutions that can provide tangible, measurable increases in efficiency to existing performance management processes through increased intelligence.
Excellent points Steve! I think those should definitely be highlighted in this space, maybe even more so than the other “sexy” stuff. I think these things are much more realistic and believable to clients.
Send some of your clients over and ask them to join the conversation!
These real-time analytic tools aren’t cheap, so we clearly have to justify the value to the business. The approach I’m taking is focusing on two areas.
First is to improve application performance. If application X performs better than you’re able to sell more products and services. Analyzing performance datasets from SAN to OS to DB to middleware to CE allows the operations teams to better understand and resolve performance problems. The advancements in these tools finally make it possible to do this in real-time.
Second is to reduce the amount of time spent adjusting thresholds manually for performance metrics. The self-learning analytical tools improve the threshold management and the result is time saved.
ManageEngine OpManager is an award winning network monitoring software that helps administrators discover, map, monitor and manage complete IT infrastructure and it is a network software that offers combined LAN, WAN, Server and Application monitoring.
http://www.manageengine.com/network-monitoring/