Becoming the next Uber is only possible when ideas can be brought to end users quickly. Some aspects of DevOps are perfect for this, though we know its success is dependent upon (surprise!) development and operations working closely together. What does this mean for developers? Delivering code faster with much higher risk? Surely there’s a way to mitigate.
Perhaps we need to rethink the metrics we’ve historically collected and begin the shift to Key Technical Metrics that may better reflect the efficacy of this new way of doing work; When tracked at each workstation all the way through CI and into Ops, metrics such as memory usage per user or request, number of SQLs, number of service calls, or transferred bytes (to name a few) can dramatically improve the speed and accuracy with which we develop. –And don’t forget the Business: How often is a new feature really used? What does it cost to run it? Let these metrics act as quality gateways and stop builds early before they crash your system: faster than ever.
In this session, we’ll look at how companies like Facebook and CreditOne apply metrics-driven DevOps. We look at use cases that crashed rapid deployments, identify metrics that identify the reason for the crash, and learn how to use these metrics to steer your pipeline to build better code, deploy faster, without failing faster.