In this podcast Clare Liguori, Principal Software Engineer at Amazon Web Services, sat down with InfoQ podcast host Daniel Bryant and discussed: the implementation of continuous delivery at AWS, the use of automation and deploying to multiple test environments, and the benefits of canary releasing.
Key Takeaways
At Amazon, a typical continuous delivery pipeline has four major phases: source, build, test, and production. For a particular microservice, there might be multiple different pipelines that are deploying different types of changes, e.g., application code, infrastructure code, operating system patches etc
Every change that goes out to production at Amazon is code reviewed, and pipelines enforce this. With automated “full CD” there are no human interactions after a change has been reviewed and pushed into the source code repository before it gets deployed to production.
AWS Launches Amazon DevOps Guru This item in japanese
by
Amazon Web Services (AWS) recently introducedAmazon DevOps Guru, one of several new machine learning-driven services. DevOps Guru detects operational issues, generates reports and notifications, and offers insights and recommendations on how to take action.
DevOps Guru is a fully-managed service that is trained to analyze logs, metrics, and events across 25 AWS resources. The service looks for behavior that deviates from patterns established by history extracted from Amazon and AWS. Users configure DevOps Guru with a list of resources to monitor. The service alerts users about problems and potential issues when it identifies anomalous situations, such as code releases that lead to abnormal behavior or resource utilization patterns that may lead to depletion.