DevOps is a buzzword right now for good reason—it can deliver significant benefits. The 2016 State of DevOps Report, which surveyed 1,400 IT professionals around the world, paints a picture of high-performing IT development organizations: those with multiple code deployments per day and less than one-hour lead time between code fixes and production deployment.
These organizations also deploy software 200 times more often than their counterparts. Their change failure rate is less than 15 percent and they spend 22 percent less time on unplanned work. The end result? Up to 10 to 20 percent less rework, which can save even small organizations millions of dollars.
But DevOps can’t achieve its potential unless there’s a tight feedback loop across all phases of application delivery.
Create a Virtuous Circle With Data
When your organization implements a continuous integration and continuous delivery (CI/CD) process, your DevOps teams must glue together a complex tool chain—one that spans requirements gathering, code management, module integration, and unit and integration testing and delivery. But all too often, these tools are used in isolation and with little measurement of their effectiveness. This ends up creating silos of disjointed information.
As the adage goes, you can’t manage what you don’t measure. You only can get the full benefit of DevOps by incorporating feedback across the different stages of the build pipeline that is based on actual data, not anecdotes, ad hoc efforts or gut feel.
Without measuring and consolidating your DevOps process data, it’s impossible to track progress throughout the CI/CD pipeline. Your team won’t be able to flag errors or report the status of defect fixes and quantify developer activity. And by giving your developers, IT operations and managers access to test data, you can tighten the feedback loop between production and deployment. You also can identify the effectiveness of test coverage and measure the quality and productivity of individual developers.
When you consolidate build pipeline and other data into a single platform, you gain end-to-end visibility into the activity and progress across the DevOps tool chain. Your developers have access to information that can help them make decisions that benefit the business through faster deployments, better software quality, improved security and less rework.
How FamilySearch Achieves 900 Deployments a Day
FamilySearch, a nonprofit family history organization with the world’s largest collection of genealogy records, uses DevOps for its CI/CI processes to deploy on Amazon Web Service (AWS). But it wanted to better track changes across its websites, so it turned to Splunk Cloud to aggregate data from across its IT environment.
FamilySearch now ingests, processes, analyzes and makes sense of up to 4TB per day of log information using Splunk Cloud. Application delivery teams built dashboards that use this data to monitor site health and the full CI/CD process. The result: FamilySearch can do 900 deployments per day with less than 20 minutes between code check-in and production release.
The Key to DevOps Success
Whether it’s a DevOps tool chain or business process, the key to improvement is the ability to record, collect and analyze data. Your developers can use a DevOps feedback loop to improve software quality, developer efficiency and release cycle time. Given the many different systems and the huge volumes of data generated across the DevOps build pipeline, eliminating the silos and blind spots in your data collection and analysis is key. To quantify DevOps, you need an effective way to ingest data from any system and format, handle large data flows in real time and provide the sophisticated data search and analysis, so your teams can easily summarize results, flag anomalies and streamline forensic analysis.