You might have heard of “The consumerization of IT,” or some bollocks like that. Well, one good example of this is social media (which we consumers love to use).

Allow me to demonstrate with a few screenshots:

End Users – The Most Reliable Alarm Clock

If I told you that 74% of IT incidents are still reported by end users first, would it shock you? Perhaps more shocking is that most DevOps or IT Operations teams aren’t paying attention to social media sentiment to detect outages, bugs or slowdowns in their applications. Yet, what better way to know if your customers are being impacted than monitoring the first place that they bitch and moan – Twitter. Yep, I think every DevOps team should be situationally aware of social sentiment, in addition to the billions of performance metrics they collect everyday.

Natural Language Processing

Sarcasm aside, it’s actually a really great use case for natural language processing and machine learning. Think about it, what if you could analyze and correlate end user events from your company’s Twitter handle with application and infrastructure events inside your data center or cloud? It’s quite a dramatic insight for IT to see the feelings of their coveted customers as they struggle to give you money, or interact with your services. If your team needs a “WTF moment” to drive change and improvement, then look no further than twitter.

Real-World Example

Well, at Moogsoft, we actually have several customers who apply this use case with Incident.MOOG every day. Most customers are used to ingesting and correlating events and alerts from their applications, infrastructure and monitoring tools. What makes Moogsoft unique, however, is our natural language processing ability, combined with our unsupervised machine learning algorithms – these make it super simple to digest, analyze and correlate information like tweets with other sentiment-related information.

Here, a Situation was created by Incident.MOOG at 21:06:06 with 1 impacted host:

The situation started with a Couchbase cache timeout event (below). Notice that 13 seconds later the application started to throw multiple Couchbase cache timeout errors:

Just one minute later, Incident.MOOG detected customer complaints from Twitter on the customer’s Twitter handle and correlated these with the Situation for IT Operations:

That is a classic example of IT Operational Analytics in action, providing automation, correlation and a unique insight so that IT Operations teams are situationally aware of what is going on around them.