Today we have monitoring tools generating logs and logs of data flowing into our data centers, letting us know everything that’s happening down to a detailed level. That’s a good thing, right? Not necessarily, because the more information we need to sift through, the more overwhelming it becomes and the more impossible it becomes for humans to track. BigPanda asserts that it can bring order to this madness and it came out of stealth today flush with $7M in Series A funding from Mayfield and Sequoia Capital to attack the problem.
BigPanda recognized that companies were getting overwhelmed with information from their IT logs, creating a problem where they couldn’t see the forest for the trees. They would come into work with screens and screens of alerts from traditional data center monitoring tools from HP and IBM to modern tools like New Relic, AppDynamics, Splunk and on and on it goes. And IT pros had to sift through all of that data making it a challenge to prioritize the flood of alerts and determining which ones truly required immediate attention. What’s more, once an IT pro found a legitimate problem that required immediate handling, it took detective work to figure out the root of the problem.
BigPanda launched the company to look for a way to simplify all of that. Here’s what they claim to do. They say they filter the firehose down to a handful of meaningful alerts. Founder Assaf Resnick said the company formed in 2012 with the intent to apply intelligence to the data pile and turn that big blob of alerts into a handful of high-level alerts that are much easier to prioritize and deal with. He said there is a lot of natural language processing and decompression going on that begins to understand the relationships between alerts and can take multiple alerts and filter them down to a single one, then learn and remember those relationships in the future.
But Resnick said they also wanted to give users enough information to solve the problem faster. By narrowing down the list of possible alerts, they wanted to give them a way to act on the most important ones more quickly then point them to the source of the problem such as a recent event like a code change that caused a chain reaction of problems.
What’s more, it includes a collaboration component, so if the problem requires outside help, you can easily pass the problem along to someone who is better suited to deal with it, and it hooks into service ticketing systems like ServiceNow and JIRA. And if a problem is something that can wait, but will need your attention at some point, you mark it to deal with later in a similar fashion to how you mark email in Mailbox to deal with at a later time, and it will surface again.
A couple of years ago I saw a talk by Paul Maritz who was CEO of VMware at the time and he said that the amount of data being generated in the data center in 2012 when he gave the speech had already surpassed human ability to deal with it. IT information had become a big data problem and that’s the case even more so today than when Maritz gave the speech. That means you have to apply data science techniques to this information to make sense of it and find those signals that really matter amongst all of the noise.
This is what BigPanda asserts it can do, and if it can truly pull it off, perhaps it can bring some order to the chaos that is the mountain of data being generated by IT monitoring tools.