Observability for Canary testing and CI/CD with Instana
Kubernetes has become a widely adopted building block of modern, agile application delivery toolchains. Orchestration technology helps IT organizations rapidly roll out new application updates and provides the platform for quickly adjusting compute capacity to manage performance.
When application delivery organizations implement agile methodologies including continuous integration and continuous delivery (CI/CD), they become capable of delivering new applications and functionalities at a much faster pace than traditional teams. To assure quality, a new testing technique becoming popular is the “Canary Testing” method, in which new software is deployed into portions of the microservice cluster, performance is analyzed, then a roll forward or roll back decision is made.
Performance and stability of Kubernetes applications
Ensuring the performance and stability of new software releases using Kubernetes becomes a critical discipline that must be considered from the beginning of a new project so that releases can immediately be identified as good or bad. If a software release is determined to be bad, then it must not be allowed to propagate to the rest of the nodes and it should be rolled back and fixed immediately. Modern application delivery organizations know that using appropriate monitoring technology is critical to this deployment and testing process.
Actapio, Yahoo Japan Corporation’s US subsidiary, realized this early on during their software delivery modernization effort. “Delivering new application functionality is critical to the continued success of Yahoo Japan,” said Norifumi Matsuya, President & CEO of Actapio in charge of cloud infrastructure engineering. “Our goal is to build the infrastructure and tools to enable our engineers to seamlessly deliver, yet precisely manage performance of, new services.”
Detection, monitoring, and classification of application infrastructure and services is challenging in highly dynamic applications that utilize heavily virtualized infrastructure. Monitoring must be fully automated, immediately responsive, and flexible enough to discover all technologies used across the application and infrastructure stack.
“Actapio wanted a monitoring solution that could work at scale in both our traditional infrastructure and our modern orchestrated environment to provide immediate performance data to our developers,” said Norifumi Matsuya, President and CEO of Actapio. “Not only did Instana meet our needs to support the full range of infrastructure and application technologies, they also meet our goals of zero-touch implementation and automation.“
As you are designing your modern application delivery pipeline, we suggest you look for the following capabilities in a monitoring platform:
- Automatic and continuous discovery of infrastructure and application components
- Automatic, zero-configuration monitoring of all components
- Immediate detection of all application changes
- Intelligent determination of problems and root cause
- Expert advice on how to solve common problems
Instana is used by leading companies globally to bring unparalleled understanding and observability to their Kubernetes powered delivery pipeline. Try Instana in your environment for free.