01732 759725 42 OBSERVABILITY With $4.2 million of seed funding and a reorganised management team, Kensu is set to make waves in the emerging data observability space. James Goulding finds out more from newly appointed CEO Eleanor Treharne-Jones Quality, Gartner cited research showing that 40% of the anticipated value of business initiatives is never achieved, to a great extent because of poor quality data in both the planning and execution phases. In addition, poor quality data reduces productivity by 20% and impairs organisations’ agility and ability to mitigate risk. Since that report was written, data volumes have increased significantly, as have automation, which relies on good quality data, and the requirement for organisations to maintain data security/privacy and demonstrate compliance, all of which increase the need for data observability, as Eleanor Treharne-Jones explains. “There’s been an explosion in data, but also in machine learning and AI and the tools that we use to try and get insight from that data. Often, by the time it is reported, there’s a big disconnect or a big distance between the people who are working with the data and making business decisions based on that data and the people involved in the data’s production in the first place. “What Kensu does with our differentiated approach to data observability – we shift left as the market likes to say – is observe problems in the data at their source, within the pipelines themselves. Our agents are constantly monitoring the freshness of the data – if you’re expecting it to be a dataset produced in the last 24 hours; the completeness of the data – if it was supposed to have 100 fields coming in, does it have 100 fields still coming out?; and observing the health of that data in real time and also in context, i.e. within the application that’s producing it. “If there’s ever a problem you will know immediately, as an alert will be sent, and you will know who the right engineer is to speak to, the person who actually manages that application. This shortens the time to remediation, which is one of the most important things and one of the challenges that the industry has had. Often, it’s been easier to identify problems but much harder to identify how to fix them. Shortening that time is the huge efficiency driver of implementing data observability.” As an example of the delays companies typically have to contend with, Treharne-Jones cites the case of a company she was talking to that took 15 days to go back and find the source of problematic data identified in an executive report. “Typically, what companies have done up to this point is to focus on using scanners and crawlers to retrospectively go and observe issues with the data. It’s better than waiting for someone to say there’s a problem, but it’s still after the fact. What we’ve done is take best practices from DataOps and some of the application observability space and embed our agents directly within the application. That means that unlike everyone else we get real-time alerts, we get continuous monitoring, and the information is provided in the context of the application where the problem occurred. We believe this will be the standard that the industry ultimately follows for data observability.” Treharne-Jones points out that having good visibility into the health of your data also supports some of the broader data governance objectives and requirements that an organisation now has. “Another thing that Kensu provides along with data health observations is lineage, exactly what the path of that data was through your pipelines and through the organisation. We can pre-populate some of the data catalogue tools with that lineage and that is also a very powerful tool that organisations see value in when they implement data observability.” Last month data observability company Kensu, founded in Belgium five years ago, secured $4.2 million of seed funding to accelerate its expansion in EMEA and the US, and appointed Eleanor Treharne-Jones as CEO to establish the company as a major disruptor in the data observability market – an appointment that has the added benefit of freeing up company founder Andy Petrella, now Chief Product Officer, to focus on continuing to develop the Kensu data observability platform. Data Observability addresses the ageold problem of poor quality, out of date, incomplete and inaccurate data and the time that data engineers waste trying to fix data quality issues. More than a decade ago, in its October 2011 report Measuring the Business Value of Data Introducing Kensu Eleanor Treharne-Jones
RkJQdWJsaXNoZXIy NDUxNDM=