I’m just back from a trip to Australia where I conducted a number of Release Management seminars. There was tremendous interest in release management in general and an even more overwhelming interest in release automation. A Delivery Manager from a large international insurance company approached me after one of the seminars. We engaged in conversation about their release management process and discussed the increasing complexity of application tiers and infrastructure environments.
This inevitably led to a white board where he proceeded to sketch their release process and draw attention to a number of their challenges. I listened intently as I’m a firm believer in the principle that prescription without diagnosis is malpractice. We quickly agreed on what appeared to be all too familiar problems:
- The increasing volume and velocity of application releases.
- The translation of development scripts and guidelines into deployment tasks and activities for pre-production and production.
- The manual sequencing of these tasks and activities across application tiers and infrastructure environments.
The following day I was invited to visit the Delivery Manager at the insurance company location. Sadly, it was announced that an application outage had occurred shortly before we arrived, and while we continued to understand the complexity and extent of their release processes and activities, it became apparent quickly that a DBA operation had been incorrectly applied, before the application was deployed. Troubleshooting an issue through a variety of largely manual deployment scripts is hard without an automated audit trail, and worse still, the automated backup recovery had failed to work also, thus causing a longer outage than would have been the case.
Their problem essentially became ours to help resolve. And now together we are working to automate their release management process, simplify the complexity, eliminate handoff tasks and assignments, and automate service tasks while eliminating errors and optimizing the integrity of application-centric deployments. Just another day’s work on the road!
