XebiaLabs this week released version 3.8 of its Deployit application release-automation platform, with automated resource discovery and on-demand auto-rollback among the software’s new features.
The Boston-based company, founded in 2008, is benefitting from the rise of DevOps, since CEO Coert Baart said that Deployit starts where build systems stop, before applications go into deployment. “The trends in the application release-automation market—continuous delivery, DevOps, virtualization, agile, compliance, high availability—have brought the market around to us,” he said. “Our first two quarters were OK, as planned. But our last two quarters are booming.”
Deployit, Baart explained, is based on four pillars: best-practice extensibility, scalability, being agentless and cloud-ready, and reporting. The first stems from plug-ins that XebiaLabs has created for all main Java EE and .NET middleware, so the same practices can be used regardless of the technology choice. Scalability comes from model-based automated generation of workflows; no manual creation is necessary.
Among the new features, automated resource discovery enables users to find a target environment, so the application can be deployed to it, said Andrew Phillips, XebiaLabs’ vice president of products. This discovery capability also can be used to audit and analyze what you have, he said.
“You can use it to build an app package that you can deploy to a new environment. You can filter out bits you want and don’t want,” he said.
The automated rollback feature is designed to give people confidence that they can return to a safe state if something goes horribly wrong. The tool “will build out a rollback sequence based on what you’ve done in the first place,” Phillips said. He admitted, though, the feature might not be used much in real-world scenarios, as developers tend to fix broken builds and move forward.
“In fact, if you build an app and it doesn’t work, you need the broken version to see how it’s broken,” he said. In QA, he noted, deployments work frequently enough where they don’t need to roll back to prior versions. Deployit will track changes as they’re made, which allows the software to automatically create a plan to undo the steps and return to the previous accepted state.
Also new in the release, Phillips said, is the ability to handle states with triggers that activate when a task or step state changes. “We’ve created a simple way to say, ‘Whenever any step is reached, we can carry out an arbitrary set of actions, such as an e-mail alert.’ Or, you can say if the step fails, try it three more times before alerting,” he explained. It also can be used to implement any policy, such as how to handle errors.
Finally, a new REST API supports continuous delivery, so tools such as Jenkins can be used alongside Deployit, which Phillips also said has undergone a refresh of its interface. “We want something in the hands of operations people supporting hundreds of apps in multiple languages and environments,” he said, “so we needed to create something easy, without 500 tabs and 500 settings that try to reflect the scale of problems.”