March 2018 Summaries

Octopus Deploy, initially focused on automating deployments within the Microsoft ecosystem, has evolved significantly since its early days, starting with version 1.5.0.1645 in March 2013, which supported deployments to Windows Azure Cloud Services. Over the years, it expanded to include deployments to Azure web apps, leveraging the popularity of Microsoft's Platform as a Service (PaaS). In June 2015, Octopus 3.0.0 introduced Azure accounts and new deployment targets, along with the capability to execute Azure PowerShell scripts, marking a shift towards first-class steps for enhanced flexibility across environments. As Octopus Deploy prepares to launch the next generation of Azure deployment support and infrastructure automation, it promises exciting new features and targets, encouraging users to follow the ongoing blog series for updates.

In a special double-episode of "Will it Deploy?", Rob Pearson explores the automation of deploying a multi-tenant SaaS web application to various customers using Octopus Deploy and AWS virtual machines. The featured application, Random Quotes, is a Microsoft ASP.NET Core 2.0 web app utilizing Entity Framework Core 2.0 and Microsoft SQL Server 2017, designed to demonstrate the complexities of multi-tenant deployments with features like customized settings and dynamic module loading. The deployment process involves several steps using Octopus Deploy, such as copying database scripts, executing migration scripts, and deploying specific modules based on customer subscriptions, all detailed in the episode's GitHub repository. Pearson invites viewers to suggest future topics and encourages subscriptions to their YouTube channel for more content.

Octopus has introduced a new feature allowing users to schedule recurring deployments, addressing a popular demand from users who wanted to automate repetitive tasks. Previously, Octopus could only schedule deployments for specific future times, but now it supports recurring deployments, which is beneficial for teams managing multiple environments like Development, Testing, and Production. This feature helps teams automate the promotion of software from one environment to another, such as from Development to Testing, ensuring stability and consistency for QA teams who need the latest version daily. Users can schedule deployments using a variety of options, including daily, weekly, monthly, or by custom cron expressions, with considerations for time zones, especially those with daylight-saving time. The feature will be available in Octopus version 2018.4.0, set for release in early April, and offers flexibility in either promoting the latest existing release or creating a new release for deployment.

Serverless computing, highlighted by services like AWS Lambdas and Azure Functions, is increasingly being seen as the future of application development, allowing users to pay only for actual code execution while supporting the creation of loosely-coupled services. At Octopus Deploy, there's an anticipation of providing robust support for AWS Lambdas shortly, and existing Azure Functions support is integrated through steps similar to those used for Azure Web Apps. The process of building and deploying Azure Functions with Octopus involves creating a simple Azure Function project in Visual Studio, packaging it using Octopus CLI, and deploying it with a Blue/Green deployment strategy to manage environments effectively. This approach utilizes AppSettings for configuration management and leverages existing Octopus functionality to switch deployment slots, thereby ensuring seamless updates and rollbacks. The underlying infrastructure of Azure Functions is based on Azure Web Apps, making it compatible with Octopus Deploy's current features, and future plans include expanding support to AWS Lambdas, encouraging the exploration of serverless computing solutions.

In October 2018, Octopus 4.0 introduced a revamped Variable Editor to address user concerns and improve functionality, including a more intuitive interface with features such as automatic variable grouping, easier text selection, and in-place scope editing. However, initial issues like non-resizable columns and slow scrolling performance prompted subsequent updates, including column resizing and performance optimizations using virtualization techniques. These enhancements aimed to improve responsiveness and user interaction, particularly in large data sets. The Variable Editor also adopted a new model allowing multiple values for a single variable across different scopes, a change that initially confused some users but eventually led to a more organized display. Additional features included handling empty variable warnings and integrating AWS account support. Looking ahead, plans for a "Variable Unification" project involve applying the Variable Editor to Tenant Variables and revising variable templates to enhance user confidence in deployment specifications.

Octopus Deploy facilitates the use of reusable PowerShell script modules for deployments across projects, enhancing automation on both Windows and Linux systems, although the latter requires additional steps for integration. Despite the fact that PowerShell script modules automatically become available on Windows, they do not on Linux, prompting the need for a workaround to achieve similar functionality. The solution involves retrieving the contents of a script module, saving them to a file, and then importing them into a PowerShell script within a bash script. This method is exemplified using a script module called "Say Hello," which demonstrates how to execute PowerShell commands on Linux, ensuring that users can effectively reuse their library of PowerShell scripts across different operating systems.

In March 2018, Octopus Deploy 2018.3 introduced significant enhancements aimed at improving infrastructure as code capabilities, which facilitate deployment and infrastructure automation. Key features of this release include the integration of GitHub repositories as a feed type, allowing for direct deployment of external resources from source control without a pre-build step, and first-class support for Terraform with new steps for applying and destroying Terraform templates. The update also introduced a deployment variable preview feature, enabling users to view variable values in specific deployment contexts, and enhanced logging capabilities with highlighted messages and artifacts that improve task log visibility. The release notes emphasize the importance of upgrading existing Octopus Deploy setups, with specific guidance for Bamboo users to update their Octopus CLI tool. Overall, these improvements aim to streamline deployment processes and enhance user experience within the Octopus ecosystem.

In March 2018, Octopus Deploy introduced a new feature allowing users to utilize GitHub as a feed source for deployments, enabling direct deployment of files from GitHub without intermediate build steps. This functionality allows Octopus to treat GitHub resources as packages by reading and parsing repository tags, which are then used to deploy scripts and templates, thus simplifying the deployment process and enhancing version control. The approach eliminates the need for packaging scripts separately, streamlining the continuous integration pipeline and facilitating better management of deployment dependencies. Despite the current focus on tags and releases, there are future plans to offer more comprehensive Git-as-a-feed support, integrating concepts like branches and commits more thoroughly, while addressing the limitations related to external hooks for triggering automatic release creation.

The blog post by Lawrence Wilson explains how to use Octopus to monitor and manage Azure resource costs by setting up notifications for cost spikes that exceed a predefined limit. Utilizing Octopus's ability to authenticate with various cloud platforms, users can deploy scripts to retrieve consumption usage details from Azure and calculate the costs associated with each resource group. A fictional company, OctoFX, is used as a scenario to illustrate the process of tagging resource groups with a NotifyCostLimit, allowing Octopus to send Slack notifications if costs surpass this limit. The post guides users through setting up an Azure Service Principal and Slack account for integration, creating a project in Octopus to automate this cost monitoring, and scheduling the deployment to run on a recurring basis. Key limitations include the exclusion of Service Manager resources and potential delays in data availability, while the script focuses on Microsoft Azure but can be adapted for AWS.

In this episode of "Will it Deploy?", Rob Pearson demonstrates the process of automating the deployment of a Microsoft SQL Server database using a Data-tier Application Package (DACPAC) to an Amazon Web Services (AWS) virtual machine via Octopus Deploy. The episode features the deployment of a simple ASP.NET Core 2.0 web app called Random Quotes, developed by marketing manager Andrew, to illustrate database changes and updates. The deployment is accomplished by using specific Octopus Deploy steps, including copying database scripts, deploying DACPAC to the SQL Server database, and deploying the ASP.NET Core application to IIS on a Microsoft Windows Server 2016 running on an AWS EC2 virtual machine. The episode's GitHub repository provides all necessary resources, and viewers are encouraged to suggest new technologies for future episodes and to subscribe for regular updates.