September 2018 Summaries

Matthew Casperson's article discusses a method for conducting UI tests on applications post-deployment using Octopus and AWS. During the NDC conference in Sydney, a question arose about running tests against previous deployments without them being part of the deployment process. The article provides a detailed walkthrough of deploying a CloudFormation template to AWS, creating an EC2 instance running Tomcat, and saving the public IP address as a variable in Octopus. This variable can be used later for UI tests by creating a separate Octopus project that utilizes a PowerShell script to access deployment details and perform network tests. Casperson highlights the flexibility of running these tests in any environment and suggests enhancements like using worker pools for international testing, scheduled triggers for automation, and notifications for error reporting.

Amazon's Elastic Container Services (ECS) offers a streamlined solution for orchestrating Docker containers, serving as a significant alternative to Kubernetes. With the release of version 2018.8.0, Octopus Deploy enhances its capabilities by supporting ECS deployments, leveraging its versioning and variable management features. The update also introduces Amazon's Elastic Container Registry (ECR) as a first-class feed type, facilitating integration despite ECR's unique authentication process requiring frequent credential updates. The post details a deployment process from Dockerfile creation to ECS deployment, using a sample HTML website to illustrate the steps involved. The integration allows users to manage image versions and variables through Octopus while using multi-package script steps to update ECS tasks and services. Although Octopus lacks an ECS-specific deployment step, its existing features enable effective deployment management. The discussion also touches on potential enhancements for ECS support in future Octopus updates, emphasizing the tool's role as a comprehensive deployment solution within cloud environments.

Octopus Server is set to introduce a long-term support (LTS) program to provide more stable and predictable release options for users. The LTS releases will occur on a three-month cadence and include critical bug fixes and security patches for up to six months without new features or minor enhancements, which will be incorporated into subsequent LTS releases. This move divides the release strategy into two lanes: the fast lane, offering the latest features and quick updates, and the slow lane, focusing on stability with long-term support. Octopus Cloud customers will automatically follow the fast lane to receive immediate updates, while self-hosted customers can choose between the two lanes based on their preference for stability or cutting-edge features. The LTS program aims to allow users to make informed decisions about their upgrade paths while maintaining a high-quality and secure product.

Octopus 2018.8 introduces new features to streamline Kubernetes deployments, allowing teams to deploy applications using Octopus environments, dashboards, and integration with other platforms. The blog outlines a comprehensive guide to configuring service accounts and namespaces with least privilege principles, deploying a web server in Kubernetes, performing blue/green updates, and using Helm for package management. It emphasizes creating a functional multi-environment cluster with repeatable deployments that scale with growing applications and teams. The guide also highlights the importance of isolating resources within Kubernetes through namespaces and service accounts, managing deployments with blue/green strategies to ensure zero downtime, and leveraging Helm for deploying Nginx Ingress Controllers across environments. The post concludes with insights on using Octopus to perform Kubernetes deployments, promoting changes from development to production, and integrating with Helm for enhanced deployment processes.

Octopus Deploy has enhanced its support for integrating Docker and Docker-Compose into continuous deployment pipelines, allowing users to leverage environment variables and image versioning through new scripting features. This integration facilitates the deployment of containerized applications by enabling dynamic variable substitution and image management directly through Octopus, thus standardizing build environments and ensuring consistency across different stages. Users can now manage Docker-Compose tasks more effectively by employing a separate .env file tailored for deployment, which can be dynamically populated with Octopus variables, thereby maintaining the flexibility to run and test locally. This approach is exemplified by a project that builds a Python web application with a Redis data store, demonstrating how Octopus' new multi-package selection and scripting capabilities can streamline the deployment process while allowing for easy updates and scalable configurations.

Michael Richardson discusses new enhancements in Octopus Deploy, specifically the ability to reference multiple packages, including container images, in script steps. Previously, users could only reference a single package that contained the script to be run, but now they can also reference packages that do not contain the script, which facilitates more flexible interactions with files and metadata. This update addresses the limitations of the previous "wrapper" package and "deploy-the-packages-first" patterns, making deployment processes less complicated and more versatile. The integration of container images allows users to capture image versions at release creation, and these can be acquired on execution targets or not at all, optimizing resource use. Using an example from Octopus HQ, Richardson illustrates how the enhancements simplify the deployment process by allowing packages to be extracted or left un-extracted as needed, and how variables from package references can be used effectively in custom scripts. Additionally, the update includes file variable-substitution features, further enhancing the flexibility and functionality of script steps within Octopus Deploy.

Octopus Deploy's 2018.8 release introduces several enhancements to script steps, including the ability to easily consume and manage multiple package references, support for variable substitution, and the capability to reference container images. The release also marks the initial introduction of Kubernetes support, available behind a feature flag to gather user feedback, though it is not recommended for production environments yet. Additionally, a new AWS Elastic Container Registry (ECR) feed type has been added to streamline the configuration of AWS credentials for accessing container images. Users are advised to follow the usual upgrade steps for Octopus Deploy and refer to the release notes for comprehensive details.