September 22, 2020
When I took a contract to deliver case studies for Cribl earlier this summer, I wasn’t looking for longer-term employment. I knew the LogStream product was well-liked, and I always enjoy the process of interviewing customers, pulling on the threads of the discussion to uncover a compelling story about their experiences – so a contract to do a few of these sounded like a good way to distract myself from being trapped in the house all summer.
When I got on the call to interview the team at Autodesk, I was already aware (I do my homework!) that they were using Cribl in a few different ways. The first thing they described to me was the major cost reduction they were able to achieve by de-duplicating data from their various logging platforms–a core Cribl use case. Once they’d far outpaced their goals for reducing duplicate ingestion, they then turned their attention to making the remaining data more valuable. With Logstream, they are able to refine and enrich the content of data for their internal customers by (re)structuring it as needed, removing useless fields, and adding actually useful fields from external sources. Lastly, they can do this any time without having to restart any collectors, indexers, or agents. No waiting for access to new, better data! Awesome, and definitely something I knew my friends in various IT practices would appreciate.
But it was this next use case that really inspired me: Using LogStream, the Autodesk team can easily preview and scrub data sources before they enter the data pipeline. As a result, they now onboard new sources very quickly. In addition to reducing cost and improving data quality, they also freed up something else: themselves. Instead of having to constantly fight fires, they’re able to work more strategically with the teams they serve to optimize and fine-tune the data flow, and are developing a plan to allow them to self-service onboarding their own sources. We started to talk about what that looks like, and I felt like I was seeing Something Major happen.
The things they’re planning to do will make it easier for everyone at Autodesk to get the data they need to drive whatever part of the business they’re in. This kind of data democratization potential is why I joined Splunk back in 2007, and why I more recently spent a couple of years evangelizing the same ideals in the form of observability at Honeycomb. So when the Criblanians suggested I consider a more permanent role with them, I didn’t have to think about it too long.
Many tools have sprouted up in the time since Splunk first began to make it possible for sysadmins to do more than grep their logs. By now, every part of the enterprise knows the value of data and is aiming to generate, retain, and analyze it–but most organizations still have to make deep compromises to get the data needed to just operate, let alone expand their effectiveness. These teams are constantly having to function within multiple suboptimal, overlapping constraints: they must constrain the data they collect, constrain access to data for compliance reasons, constrain the amount of data they index and store for cost reasons–all of which limits their ability to fully understand the problems they face or leverage any advantages they might have.
Cribl is offering something that makes all of this much less of a zero-sum game. Instead of having to choose one path for the data, one use case to optimize for, one analytics platform to send it to, one retention policy, one storage destination–LogStream turns these constraints into “Yes, and” choices–enabling greater access to all types of data (when so many vendors want to lock in their users and constrain their choices). So–yes to the expansion of data democracy, and yes–I’m enthusiastically signing up for more here at Cribl! See you in our user community Slack \o.