Sharing the Lessons Learned During Loggly Gen2 Beta

 

A lot of planning and effort went into the second generation of Loggly. We knew that the stakes were high if we wanted to grow our position as the world’s most popular cloud-based log management service. And we knew that real-world feedback from real DevOps people —both the good and bad—would be the key to our success.

The Loggly Gen2 beta was an iterative process with multiple phases of private and public beta. Early on, we proactively recruited a wide variety of our Gen1 customers. We reached out to our most active customers, the people who live and breathe Loggly and process the largest data volumes. But we also recruited customers who were having problems with Gen1 because we knew that understanding their struggles would help us build a much better service. We included new customers on a rolling schedule so that we would be able to get closer to each participant and have time to understand their experiences.

Since all of the beta participants were already using Loggly, the Gen2 beta benefited from a side-by-side, real-world experience. When we received participants’ log data, we made it available simultaneously in both Gen1 and Gen2 systems. A customer with a burning production issue could do root cause analysis in Gen1 and then switch over to Gen2 to compare the troubleshooting experience or dig deeper into trend monitoring. So: even when Gen2 had not yet met our standards for a “production-ready” service, we were able to see how people used it to solve actual operational problems, and do so in real-time.

The Loggly Gen2 beta pointed us to some features that would be big winners; for example, the grid functionality and trending.

Quick insights with trends

The time to bake with real customers, real data and 1-to-1 interviews also pinpointed the specific features and areas that would need to evolve early after Gen2 general availability.  We’re now a handful of releases post launch and the time invested keeps paying off.

Performance and scalability were critical requirements for Gen2, and our beta gave us a way to put our new architecture to the test, ahead of true full customer load. Loggly itself was a Gen2 beta participant, and the combination of all our log file data, our simulated load, plus our high volume participants taxed the system earlier than our engineering team had anticipated. We suspended the beta for two weeks to address the issues, and the end result was a service that has performed phenomenally even as signups, activations, and usage have exceeded all of our goals and projections.

We now have thousands of customers on Loggly Gen2, and our true reward is knowing they are getting value from the enterprise class product they deserved and one that has been properly battle tested.


Share Your Thoughts

Top