With containers rapidly becoming the de facto standard for application delivery in the cloud, traditional CI/CD methods can bottleneck software delivery performance. It requires a stronger unification of developer and operations workflows with cloud-native tooling.
When developers and operations teams collaborate during the build phase of an application, bugs and issues are found sooner. There’s less wasted effort throughout the DevOps process.
Addressing an increased focus on managing and orchestrating containers
Adapting continuous delivery for cloud-native workflows doesn’t mean abandoning existing DevOps toolchains and practices. Common IT processes and tooling like configuration management systems, log aggregation systems, and monitoring systems don’t disappear: these systems evolve and adapt to cloud-native requirements.
Configuration management systems are still used to deploy and manage the life cycle of core distributed systems, such as the container orchestration system, its delivery pipeline, and other dependencies (like data management systems that aren’t running in containers).
In a cloud-native workflow, there’s still a focus on log aggregation systems to provide data for audits, forensics, and predictive analytics. Operations teams must now intake and parse logs that come from the container orchestration system and the container delivery pipeline.
Evolving the infrastructure delivery pipeline with Dev and Ops
Another critical component is the container delivery pipeline. These systems and processes automate the build and test workflows for every code check-in. They also automate the deployment of successful check-ins into the container orchestration system.
The biggest shift in the operations workflow is to move core aspects of the infrastructure delivery life cycle, such as remediating vulnerabilities within infrastructure, out of the runtime monitoring of production systems and into the build pipeline.
For example, instead of patching a vulnerable package on the running container, the ops team needs the ability to flag a vulnerable package version using container inspection tools, trigger a rebuild of the container image, scan the image for vulnerable packages as part of the CI/CD pipeline, and only deploy images that pass these scans from a build perspective.
This shift from reacting in a runtime scenario to proactively remediating issues within every iterative build requires a continuous delivery solution with a cloud-native focus. This solution should be a bridge in unifying workflows between Dev and Ops teams.