January 2021 Summaries

Serverless computing represents a shift away from managing traditional server infrastructure, with platforms like AWS Lambda allowing developers to focus on code execution without worrying about server management. This approach is particularly advantageous for infrequent or scalable workloads, such as functions triggered by specific events. The blog post discusses two deployment strategies for serverless applications: self-contained and decoupled. Self-contained deployments bundle all related resources into a single CloudFormation stack, simplifying management but limiting flexibility, as all components are tightly coupled. On the other hand, decoupled deployments allow independent management of each Lambda function while sharing a common API Gateway, which provides more flexibility for microservices architecture. The post highlights the challenges of using AWS's Lambda aliases and API Gateway stages for managing environments and deployments and suggests a model where each environment corresponds to a single API Gateway and stage, while multiple Lambdas can deploy independently. This approach supports common deployment patterns like hotfixes and feature branches, offering a scalable solution for complex serverless architectures.

Infrastructure as Code (IaC) is an approach that allows teams to automate the creation of infrastructure resources using declarative configuration, and Terraform, an open-source tool by Hashicorp, is a prominent solution in this field. The introduction of the official Terraform provider for Octopus Deploy marks a significant development, allowing users to manage and configure Octopus Deploy instances, both on-premises and in the cloud, through Terraform scripts. This provider, initially a community initiative, benefits from being officially supported, ensuring continuous updates, improved testing, and new features. With Terraform's capabilities, users can provision, update, and manage infrastructure configurations efficiently, leveraging the benefits of automation and version control. The new provider, powered by a cross-platform client in Go, simplifies interaction with Octopus without direct REST API calls, enhancing the integration process. This IaC approach is distinct from Config as Code, which focuses more on automating processes within projects while Terraform encompasses a broader scope, managing infrastructure, projects, and system configurations. The official provider is available on the Terraform Registry, and its use is supported by extensive documentation and community contributions.

The Tentacle agent's new support for ARM and ARM64 hardware enables deployment of applications and services to devices like Raspberry Pi 3 & 4 and AWS A1 EC2 instances, offering a more secure alternative to SSH by avoiding the need for open firewall ports. This update facilitates running Tentacle on ARM servers, which are cost-effective and offer faster compute capabilities compared to x86 servers. The guide provides a detailed walkthrough for configuring Tentacle on Linux using a Raspberry Pi, explaining how to connect it to an Octopus Cloud instance and manage it through Runbooks for application updates and system management. The Polling Tentacle further simplifies deployments by establishing outbound connections to the Octopus Server, thus bypassing complex firewall settings and enhancing security in sensitive environments.

Marcel Kesselring discusses the process of using the octopus-deploy-plugin, a Gradle plugin developed by Liftric, to facilitate package uploads to Octopus Deploy for projects not supported natively by build servers. The plugin supports creating and uploading package build-information, uploading packages, and generating build-information from git commits. A sample Spring Boot project is used to demonstrate the plugin's functionality, emphasizing the use of Kotlin for Gradle build files and incorporating necessary dependencies like Spring Web and Spring Boot Actuator. The configuration process includes setting up the plugin with Gradle's Lazy Configuration approach, specifying project details, and securely handling API keys, with Liftric recommending tools like Hashicorp Vault for secrets management. The plugin allows users to push build information and packages to Octopus Deploy, with notes on adapting file naming conventions to meet Octopus Deploy's requirements. Further details and the full feature set are available in the plugin's repository.

Alex Yates discusses the contentious use of the job title "DevOps Engineer" in the tech industry, arguing that it misrepresents the collaborative and comprehensive nature of DevOps practices. He points out that DevOps is not a task for a single individual or team but rather a broad organizational approach involving multiple roles and departments working together. Yates suggests that assigning the title "DevOps Engineer" can lead to misunderstandings about DevOps, reducing it to mere deployment pipelines or infrastructure as code, which overlooks its broader cultural and procedural impacts. Instead, he advocates for more precise job titles such as "Build/Release Engineer" or "Infrastructure Engineer" that reflect specific responsibilities. Yates also emphasizes the importance of recognizing DevOps as an organizational capability rather than an individual one, and highlights the potential negative perception of companies that use "DevOps Engineer" as a job title.