One Step Beyond
In a recent interview that TNS founder Alex Williams had with Kubernetes advocate Kelsey Hightower, the always-eloquent Hightower noted that “We hear this every time there's a technology transition: people believe that the shoulders that we stand on to get to where we are now ... can go nameless," Hightower said.
Hightower was discussing, along with Canonical CEO Mark Shuttleworth, how Kubernetes relies on Linux, even as people forget how useful Linux has been, thanks to its maturity and reliability, in the Kubernetes-fueled container revolution. And in time, they hope, developers would forget about Kubernetes too, even as they increasingly relied upon it.
“If you can build something that people are so comfortable with that they don't worry about it anymore, you're essentially freeing all of that intellectual energy up to then focus on problems that arguably are more interesting,” Shuttleworth said, in this upcoming edition of The New Stack Makers podcast.
Building on the innovation of others is the natural progress of technology. For an upcoming ebook that we are putting together for LogDNA, “Cloud Native Observability for DevOps Teams,” we got an opportunity to discuss with LogDNA cofounder Lee Liu about what inspired him to start the company when so many other competitive businesses already crowded the field.
In a nutshell, the company needed logging to scale to the next level, to build on what others have done, and at the same time allow developers to forget about the previous building blocks. In its original (since discarded) business plan, LogDNA was more than content to use Logstash for the ingestion of the logs that needed to be analyzed, but the problem with Logstash, as Liu explained, was “you need to tell it what type of logs you're ingesting so it can do the regex filtering and to enrich data and all that stuff.” That was fine if you were monitoring one app, say, but how about the whole stack, including the database, app server and network drivers? The job would quickly spiral out of control.
“So, we wrote our own injector that basically would take the data that the logs are coming in, and will auto-detect what type of file it was, and auto-parses that,” he explained. And thus LogDNA was born, building on the ideas of log analysis, while at the same time, taking them one step beyond.