Over the last several months I have been working in Australia on an enterprise uplift program, moving a large organisation from expensive infrequent releases towards continuous delivery and DevOps.
As a group, we have focused on a simple ritual of experiment, measure, learn, and pivot towards the highest value.
We’ve been looking at new DevOps enabling technologies like containerisation and automated provisioning of infrastructure and run several experiments with Wildfly, IBM Liberty, traditional WAS as well as existing integration using vRealize and Puppet on RHEL.
In the beginning, what surprised and slightly deflated my enthusiasm was pivoting back to the status-quo.
We found containerisation would indeed provide huge gains within the CI process and confidence in shipping production ready infrastructure and software. But, the biggest and quickest wins were making some small but significant changes to existing processes and technologies. By hardening the CI pipeline and deploying to existing endpoints we will be able to leap forward in the number of deployments the organisation can make with confidence.
By demonstrating value we can provide the cultural momentum to make bigger changes in both technology and process. Throwing out the bath water is not always the highest value option…
How much value? Sizing our footprint from 4000 manual deploys to 20,000 automated seems like a good reason to pivot.