August 2015 Summaries

Octopus Deploy 3.1, a pre-release version available as of August 27, 2015, introduces several significant features, including enhanced reporting capabilities, Tentacle upgrades supporting TLS 1.2, and new Azure step-types. The reporting feature allows users to generate deployment history reports using external tools like Excel or PowerBI via an XML feed or a SQL database table. The update requires .NET 4.5 on the server and enables TLS 1.2 for Octopus/Tentacle communications, although older Tentacle versions can still be used. The Azure deployment process has been revamped with new step types such as Azure Cloud Service, Azure Web App, and Azure PowerShell, which can automatically migrate existing Azure steps from version 2.6 without modification. While 3.0 Azure targets remain available, they are deprecated, and users are encouraged to transition to the 3.1 step types. Additionally, a new feature allows deployments to sub-directories of Azure Web Apps, addressing user requests for more flexible deployment locations. The release prioritizes these Azure changes, with future updates planned to include features like Branching Support and Multi Tenancy.

Octopus Deploy High Availability is designed to enhance the performance and reliability of Octopus Deploy, particularly for customers who consider it a critical infrastructure component. By adding nodes to a High Availability cluster, users can expect a linear increase in performance, as demonstrated in tests conducted on Amazon EC2. These tests showed significant time reductions in deploying packages when additional Octopus Servers were used, with deployments completed in 35 minutes using one server, 18 minutes with two, and 11 minutes with three. However, some bottlenecks were identified, such as the use of an external NuGet feed and shared Tentacle deployments, which could impact performance. Adjustments, like using the built-in package repository and managing concurrent tasks, were suggested to overcome these issues. The High Availability feature was set to be available for trial soon after the tests.

Octopus Deploy is preparing to launch its High Availability (HA) edition, designed to support distributed workloads across multiple servers to enhance deployment efficiency and reduce the risk of timeouts during large-scale operations. The HA edition employs load balancing to distribute tasks across Octopus servers, each managing web requests and background tasks while sharing a SQL Server database and filesystem for task logs and packages. Performance challenges, such as hardware limitations during extensive deployments, prompted the implementation of a running task cap, allowing users to configure the maximum number of simultaneous tasks per server. This approach prioritizes deployment speed over concurrency, ensuring faster completion of critical deployments. Additionally, the running task cap facilitates server maintenance by enabling a "server draining" feature, allowing servers to be gracefully shut down for updates without disrupting ongoing tasks.

Paul Stovell, founder and CEO of Octopus Deploy, shares insights into the company's evolving work methodologies and tools that have helped maintain smooth operations. Initially adopting a Scrum-based process with bi-weekly shippable products, the team faced challenges when they went eight months without a major release, highlighting the issues of working without strict time constraints and the inefficiencies of tackling massive architectural changes simultaneously. To address these issues, Octopus Deploy implemented strategies such as setting clear deadlines, conducting regular "bug bashes" to identify and fix issues quickly, and adopting Kanban to manage workflow. The company shifted to smaller, more focused teams with shorter stand-ups, enabling them to ship updates more frequently and maintain momentum. Tools like Slack, Google Hangouts, and Waffle.io have been integral in facilitating communication and collaboration, particularly in a distributed work environment. Stovell emphasizes the importance of regularly reassessing and evolving processes to avoid complacency and maintain a strong focus on shipping quality products.