The IT market is seeing a rush of startups that monitor apps, data, websites, servers and networks. This is a great trend for best-of-breed IT shoppers, but deployment of multiple tools creates its own problems.
These monitoring tools, even if they are producing data about different IT events, may be pointing to common problems. What if someone could take all that divergent data, regardless of source, analyze it, and pull out the most important thing to know? That's what BigPanda, the latest startup in this market segment, is doing.
BigPanda's approach takes data from multiple performance monitoring sources, which in total may be sending out hundreds or thousands of alerts, and then searches out their correlations and dependencies, indicated by tags and their relationships in time, to discover the source of a problem.
Industry analysts are positive about Mountain View, California-based BigPanda's effort. "I haven't seen anything this dramatic before in terms of the automation and simplicity," said Donnie Berkholz, of RedMonk. At Enterprise Management Associates, analyst Dennis Drogseth, said what BigPanda is doing is not a new challenge, but its approach is "distinctive," and "very efficient and relevant."
The true test for BigPanda began with Tuesday's launch of its platform, which up until now has been in beta. The company began in 2012 with $1.5 million, and said this week it has received $7 million in new funding.
BigPanda uses natural language processes to take what is unstructured data from a variety of monitoring sources, couples it with analysis that identifies the potential source of a problem, such as a recent code change, and then automates notification. Its operational platform uses a SaaS environment, and it supports a growing list of third-party tools.
Assaf Resnick, BigPanda co-founder and CEO, said the company has hired data scientists to develop the machine-learning capabilities needed for the analysis system.
Resnick said the appeal of best-of-breed tools has also brought a "dramatic fragmentation" to the market, creating big data problems for users and opportunity for his firm.
Whether you want to call what's going on in system monitoring a "dramatic fragmentation" or an abundance of choice, is your call. But In the past decade this market has grown, and continues to grow, with companies such as New Relic, founded in 2008, AppDynamics in 2008, DataDog in 2010, StatusCake in 2012, BugSense, founded in 2011 and acquired last year by Splunk, founded in 2003; Scout Monitoring in 2008, Sentry in 2010 and Pingdom in 2007.
Open source monitoring tools, including Sensu, Nagios, Incinga and Zabbix, are also a major part of this picture.
Drogseth said a global company may have as many as several hundred monitoring tools, including in-house developed systems, in various data centers and geographies. He said BigPanda doesn't offer predictive capabilities, but is scalable and can mature into something more.
Berkholz said it will be interesting to see whether incident management "develops into a larger stand-alone market versus driving more monolithic vendors to incorporate similar features and processes." However, "it could go either way at this point, since the need for incident management will continue to increase as IT infrastructures get more and more complex," he said.
This story, "Startup Aims to Bring Order to Performance Monitoring " was originally published by Computerworld.